close

Вход

Забыли?

вход по аккаунту

?

Electromagnetic exposure in a phantom in the near and far fields of wire and planar antennas

код для вставкиСкачать
ELECTROMAGNETIC EXPOSURE IN A PHANTOM IN THE NEAR AND FAR FIELDS
OF WIRE AND PLANAR ANTENNAS
by
Md. Anas Boksh Mazady
Bachelor of Science
Bangladesh University of Engineering and Technology, 2006
Submitted in Partial Fulfillment of the Requirements
For the Degree of Master of Science in
Electrical Engineering
College of Engineering and Computing
University of South Carolina
2010
Accepted by:
Dr. Mohammod Ali, Director of Thesis
Dr. Grigory Simin, Reader
Tim Mousseau, Dean of The Graduate School
UMI Number: 1483674
All rights reserved
INFORMATION TO ALL USERS
The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
UMI 1483674
Copyright 2011 by ProQuest LLC.
All rights reserved. This edition of the work is protected against
unauthorized copying under Title 17, United States Code.
ProQuest LLC
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, MI 48106-1346
© Copyright by Md. Anas Boksh Mazady, 2010
All Rights Reserved.
ii
DEDICATION
To my loving parents – I cannot emphasize how lucky I am to be your son
iii
ACKNOWLEDGEMENTS
I would like to take this opportunity to express my gratitude to those people
without whom this work would have not been possible. At first, I would like to thank my
academic supervisor Dr. Mohammod Ali for his active guidance and having faith in me. I
can barely remember an occasion when I had a question and he was not there to help me
around. A number of graduate courses I took were actually taught by him that gave me
the foundation to carry out the research in this topic.
I would also like to show my gratitude to Gernot Schmid and the members of
Seibersdorf Research Laboratories for providing with the measurement data and Dr.
Grigory Simin, the reader of this thesis. I want to acknowledge Mobile Manufacturing
Forum and GSM Association for initiating this project. All the members of my research
lab had active support on me and their invaluable knowledge often illuminated my path to
success.
I am indebted to my parents for instilling the belief in me that someday I could be
their hero. I am thankful to my sisters, uncles, and aunts for being so supportive.
iv
ABSTRACT
Due to the wide availability and usage of wireless devices and systems there have
been and are concerns regarding their effects on the human body. Respective regulatory
agencies have developed safety standards based on scientific research on electromagnetic
(EM) exposure from wireless devices and antennas. The metric that quantifies the
exposure level is called the Specific Absorption Rate (SAR). Wireless devices must
satisfy the regulatory standards before being marketed. In the past, researchers have
primarily focused on investigating the EM exposure from wireless devices that are used
very near to the user’s head or body (less than 25 mm). But as time progressed many
more wireless devices have become ubiquitous (vehicular wireless devices, laptop
PCMCIA cards, Bluetooth dongles, wireless LAN routers, cordless phone base stations,
and pico base stations are to name a few) and are operated at distances greater than 25
mm yet smaller than 200 mm. Given the variations in operating frequency, distance, and
antenna size and type it is challenging to develop an approach using which EM exposure
from a wide variety of wireless devices can be evaluated. The problem becomes more
involved owing to the difficulties in identifying the antenna zone boundaries, e.g. reactive
near-field, radiating near-field, far-field etc. The focus of this thesis is to investigate a
large class of low and highly directive antennas and evaluate the EM exposure from them
into a large elliptical phantom. The objective is to be able to predict threshold power
levels that meet the SAR limits imposed by the regulatory agencies. It was observed that
v
among the low directivity antennas at close near-field distances, electrically small
antennas induced distinguishably higher SAR than electrically larger antennas. But
differences in SAR were small as the phantom moved into the far-fields of the antennas.
SAR induced by highly directive antennas were higher when the phantom was in the farfield of the antennas and was facing the antenna frontal plane. The same was not true
when the phantom was in the near-field of the antennas. Finally, by analyzing the
simulation and measurement data threshold power formulas were developed for low
directivity antennas using which power levels corresponding to the safe exposure limits
independent of device type or geometry can be estimated.
vi
TABLE OF CONTENTS
DEDICATION ....................................................................................................................... iii
ACKNOWLEDGEMENTS........................................................................................................ iv
ABSTRACT ............................................................................................................................v
LIST OF TABLES .................................................................................................................. ix
LIST OF FIGURES ...................................................................................................................x
I. INTRODUCTION ..................................................................................................................1
MOTIVATION AND OBJECTIVE 1.1 ...................................................................................1
EXPOSURE STANDARDS 1.2.............................................................................................4
IDENTIFYING REGIONS 1.3 ..............................................................................................4
SELECTION OF FREQUENCY 1.4 .......................................................................................5
COMPUTATIONAL METHOD 1.5 .......................................................................................7
PHANTOM GEOMETRY AND PROPERTY 1.6......................................................................9
ANTENNA TO PHANTOM DISTANCES 1.7 .......................................................................11
II. SAR INDUCED BY DIPOLE ANTENNA .............................................................................13
ANTENNAS STUDIED FROM DIPOLE CLASS 2.1 ..............................................................13
FREQUENCY OF OPERATION 2.2 ....................................................................................14
DIPOLE ORIENTATIONS 2.3 ...........................................................................................15
SIMULATION SETUP 2.4.................................................................................................16
MEASUREMENT PROCEDURE 2.5 ...................................................................................17
RESULTS 2.6 .................................................................................................................21
CONCLUSION 2.7 ...........................................................................................................26
III. SAR INDUCED BY MONOPOLE AND PLANAR INVERTED-F ANTENNA (PIFA) ...............27
MONOPOLE ANTENNA MODELS 3.1 ..............................................................................27
PLANAR INVERTED-F ANTENNA MODELS 3.2 ...............................................................28
SIMULATION SETUP 3.3.................................................................................................29
RESULTS 3.4 .................................................................................................................30
CONCLUSION 3.5 ...........................................................................................................39
vii
IV. SAR INDUCED BY PATCH ANTENNA, DUAL-BAND PIFA, AND DUAL-BAND IFA ........41
MICROSTRIP PATCH ANTENNA MODELS 4.1 .................................................................41
RESULTS (MICROSTRIP PATCH ANTENNA) 4.2 ..............................................................42
MODELS OF OTHER PLANAR ANTENNAS 4.3 .................................................................47
SAR RESULTS WITH OTHER PLANAR ANTENNAS 4.4 ...................................................48
CONCLUSION 4.5 ...........................................................................................................51
V. SAR INDUCED BY ANTENNA ARRAYS............................................................................52
MODELS OF PATCH ANTENNA ARRAY 5.1 ....................................................................52
SAR RESULTS WITH PATCH ANTENNA ARRAY 5.2 .......................................................53
MODELS OF DIPOLE ANTENNA ARRAY 5.3 ...................................................................57
SAR RESULTS WITH DIPOLE ANTENNA ARRAY 5.4 ......................................................65
CONCLUSION 5.5 ...........................................................................................................67
VI. THRESHOLD POWER RATIONALE ..................................................................................68
THRESHOLD POWER RATIONALE AT 40 MM DISTANCE 6.1 ............................................68
THRESHOLD POWER RATIONALE AT 100 – 200 MM DISTANCES 6.2 ..............................72
UNDERESTIMATION 6.3 .................................................................................................75
THRESHOLD POWER RATIONALE FOR DIRECTIVE ANTENNAS 6.4 .................................78
CONCLUSION 6.5 ...........................................................................................................79
VII. CONTRIBUTION AND FUTURE WORK ...........................................................................81
CONTRIBUTION 7.1 .......................................................................................................81
FUTURE WORKS 7.2 ......................................................................................................82
VIII. REFERENCES ..............................................................................................................84
viii
LIST OF TABLES
Table 1.1. Frequency bands of commonly used communication devices........................... 6
Table 1.2. Electrical properties of tissue simulating liquids for homogeneous phantom . 11
Table 2.1. Measurement equipment .................................................................................. 20
Table 2.2. Computed and measured SAR for  and  dipoles ................................. 22
Table 3.1. PIFA dimensions for 900 and 1900 MHz ........................................................ 28
Table 3.2. PIFA dimensions at 2450 MHz ....................................................................... 29
Table 3.3. SAR induced by monopole antennas ............................................................... 30
Table 3.4. SAR induced by PIFAs .................................................................................... 34
Table 4.1. HFSS design data of microstrip patch antennas on FR4 substrate .................. 42
Table 4.2. SAR Comparison of PIFAs and IFAs with Patch Antennas (Conventional) .. 49
Table 4.3. SAR Comparison between Dual Band PIFAs and Patches (Flipped) ............. 51
Table 5.2. Peak Realized Gain of Different Patch Arrays ................................................ 53
Table 5.3. IE3D simulation results for a 3-element dipole array ...................................... 58
Table 5.4. SAR induced by different antennas at 2450 MHz and at d=200 mm .............. 65
Table 6.1. Lowest Pth,10g at 40 mm distance ..................................................................... 70
ix
LIST OF FIGURES
Figure 1.1. Homogeneous elliptical phantom ................................................................... 10
Figure 2.1. Dipole antenna prototypes for measurement .................................................. 15
Figure 2.2. Definition of distance for dipole antennas...................................................... 16
Figure 2.3. An ELI4 Phantom ........................................................................................... 18
Figure 2.4. Test chamber with wooden phantom holder and probe positioning unit ....... 19
Figure 2.5. Antenna holder with antenna below phantom ................................................ 19
Figure 2.6. SAR vs. frequency due to dipoles at 40 mm distance from the phantom ...... 22
Figure 2.7. SAR vs. frequency due to dipoles at 100 mm distance from the phantom .... 23
Figure 2.8. SAR vs. frequency due to dipoles at 200 mm distance from the phantom .... 23
Figure 2.9. Tissue conductivity vs. frequency plot ........................................................... 25
Figure 3.1. Monopole antenna and box model ................................................................. 27
Figure 3.2. Geometry of planar inverted-F antenna for 900 and 1900 MHz .................... 28
Figure 3.3. A surface mounted planar inverted-F antenna (PIFA) for 2450 MHz ........... 29
Figure 3.4. SAR vs. frequency of monopoles at 40 mm distance from the phantom ...... 31
Figure 3.5. SAR vs. frequency of monopoles at 100 mm distance from the phantom ..... 31
Figure 3.6. SAR vs. frequency of monopoles at 200 mm distance from the phantom ..... 32
Figure 3.7. 1-g averaged SAR distribution by 1900 MHz and 2450 MHz monopoles .... 33
Figure 3.8. 1-g averaged SAR distribution by 1900 MHz dipoles ................................... 33
Figure 3. 9. SAR vs. frequency due to PIFAs at 40 mm distance from the phantom ....... 35
Figure 3.10. SAR vs. frequency due to PIFAs at 100 mm distance from the phantom .... 35
x
Figure 3.11. SAR vs. frequency due to PIFAs at 200 mm distance ................................. 36
Figure 3.12. SAR vs. frequency of flipped PIFAs at 40 mm distance .............................. 38
Figure 3.13. SAR vs. frequency of flipped PIFAs at 100 mm distance ............................ 38
Figure 3.14. SAR vs. frequency of flipped PIFAs at 200 mm distance ............................ 39
Figure 4.1. Schematic of a microstrip patch antenna on a metal box ............................... 41
Figure 4.2. SAR vs. frequency of patch antennas at 40 mm distance............................... 42
Figure 4.3. SAR vs frequency of patch antennas at 100 mm distance.............................. 43
Figure 4.4. SAR vs frequency of patch antennas at 200 mm distance.............................. 43
Figure 4.5. SAR distribution of patch antennas at 200 mm distance (Conventional) ...... 44
Figure 4.6: SAR vs. frequency of flipped patch antennas at 40 mm distance .................. 45
Figure 4.7. SAR vs. frequency of flipped patch antennas at 100 mm distance ................ 45
Figure 4.8. SAR vs. frequency of flipped patch antennas at 200 mm distance ................ 46
Figure 4.9. Schematic of a 2450/6000 MHz Dual-Band PIFA ......................................... 47
Figure 4.10. A Dual-Band IFA for operation in the 2450/6000 MHz bands .................... 48
Figure 4.11. SAR distribution of dual band PIFA at 6000 MHz (Conventional). ............ 50
Figure 5.2. Patch array prototypes -- (a) on FR4 substrate, (b) on RO4003 substrate ..... 52
Figure 5.3. SAR vs. frequency of patch antenna arrays at 40 mm distance (conv) .......... 53
Figure 5.4. SAR vs. frequency of patch antenna arrays at 100 mm distance (conv) ........ 54
Figure 5.5. SAR vs. frequency of patch antenna arrays at 200 mm distance (conv) ........ 54
Figure 5.6. SAR vs. frequency of patch antenna arrays at 40 mm distance (flipped) ...... 55
Figure 5.7. SAR vs. frequency of patch antenna arrays at 100 mm distance (flipped) .... 56
Figure 5.8. SAR vs. frequency of patch antenna arrays at 200 mm distance (flipped). ... 56
Figure 5.9. SAR distribution for flipped patch array at 40 mm distance. ......................... 57
xi
Figure 5.10. Schematic of a 3-element dipole antenna array ............................................ 58
Figure 5.11. Gain pattern of the dominating E-field of a 3-element dipole array. .......... 59
Figure 5.12. Gain pattern of the E-field of a 3-element dipole array without reflector. ... 60
Figure 5.13. Gain pattern of the dominating E-field of a 5-dipole array with reflector. .. 62
Figure 5.14. Gain pattern of the E-field of a 5-dipole array without reflector. ............... 63
Figure 5.15. Gain pattern of dominating E-field of a 7-dipole array with reflector ......... 64
Figure 5.16. Scatter plot of SAR vs. Directivity ............................................................... 66
Figure 6.1. Threshold Power for 10-g SAR at 40 mm distance vs. -7 dB BW................. 69
Figure 6.2. Lowest Pth,10g vs. frequency at 40 mm distance.............................................. 71
Figure 6.3. Pth,10g (mW) with frequency at 100 and 200mm for /2 dipoles .................... 73
Figure 6.4. Relationship of Pth,10g (mW) with separation distance ................................... 73
Figure 6.5. Relationship of Pth,10g (mW) with directivity at d=100 and 200mm .............. 74
Figure 6.6. Underestimation using (1) at 40 mm distance ................................................ 76
Figure 6.7. Underestimation using (2) at 100 mm distance .............................................. 77
Figure 6.8. Underestimation using (2) at 200 mm distance .............................................. 77
Figure 6.9. Pth,10g vs. directivity of high directivity antennas ........................................... 79
xii
CHAPTER 1
INTRODUCTION
1.1
MOTIVATION AND OBJECTIVE
With the advent of mobile telephony during the 1960s various wireless communication
devices became widespread in the consumer market. From then it has been a growing
concern if electromagnetic radiation from those devices have any adverse health effect. In
the early days, scientists performed a number of studies on laboratory animals, such as –
rabbits, rats, squirrels, monkeys etc. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10]. The most
important conclusion from those researches was finding a correlation between the
antennas’ radiated power and disruption in behavior of those animals. As time
progressed, various devices have been studied exposing different parts of the body in
different environments. Variations of devices include base stations [11] [12], Wi-Fi
transmitters such as wireless routers or laptop antennas [13] [14], mobile phones [15]
PDA’s [16], and different antennas such as monopoles, dipoles, PIFA’s, whip antennas
[15], IFA’s etc. Variations of orientations include the antenna being placed next to the
eye [17], next to the ear etc. Variations of environments include open environments such
as air, or closed environments as elevators [18], cars, trains [19] [20] etc. Variations of
phantoms or prototypes include realistic homogeneous flat phantoms, homogeneous
anthropomorphic phantoms, and heterogeneous realistic phantoms. Those researches
1
were mostly focused on devices that operate near the body; one very common example of
such devices most used in the literature is the mobile phone. But there are a large number of
devices that neither operate in physical contact of the body nor can they be emulated with
quasi-uniform plane waves. Examples include vehicular electronics, laptop PCMCIA cards,
Bluetooth dongles, wireless LAN routers, cordless phone base stations, pico base stations etc.
Only a few studies on those devices can be found in the literature [14]. In a previous paper
[21] our group studied devices near the body, i.e. devices ranging from the ones which touch
the phantom to the ones which are 25 mm away from the phantom. At 900 – 6000 MHz
frequency range, the phantom is in most cases, if not all, in the reactive near-field distance of
those devices. But as the distance increases this statement is no longer true, especially at high
frequencies. If it was known from before that the user was in the far-field of those devices,
plane waves could be used to emulate them. Unfortunately this region separation is not a very
easy task in the modern day devices because of their complex geometries. It is often obscure
which part of the device is actually radiating in those devices. Besides there is no straight
forward rule to identify field regions of the devices that are neither as small to fit in a radian
sphere nor as large compared to a wavelength. In this paper we studied SAR of canonical
antennas, both numerically and experimentally, that are used in devices in 900 – 6000 MHz
frequency range and at 40 – 200 mm distances from a user in their normal modes of
operation. Moreover we included some high directivity antennas in our study which might be
used at this distance range and showed how SAR is affected by their directivities. We did not
study SAR beyond 200 mm distance because then other metrics such as spatial averages of
power density or the mean squared electric and magnetic field strengths are used (IEEE
C95.1-2005) [22].
2
Unlike most other works found in the literature, we used a large homogeneous
elliptical flat phantom as recommended in the IEC 62209 [23]. This is because spherical
wave fronts from antenna radiation become spatially sparse as they travel a large distance. So
a larger phantom maximizes the electromagnetic coupling. It has also been shown
experimentally that a larger phantom induces higher SAR [24].
Finally, a rationale has been investigated between SAR and antenna parameters such
as frequency, separation distance and directivity. It has been hinted that formulas can be
developed to compute the maximum value of threshold power of the antennas below which
devices can be deemed as compliant with the regulatory standards. An approach to develop
such a formula has been shown that can calculate threshold power for compliance with 10-g
peak spatial averaged SAR for low directivity antennas. The said formula is independent of
device types, sizes, shapes, structures, antenna placement on devices etc. and only depends
on parameters which should be known beforehand. The approach can be followed to develop
a robust formula which will calculate maximum antenna threshold power for devices
touching the user to devices 200 mm away from the user and for both low directivity and
high directivity antennas. Development of such a formula is important since it will greatly
simplify the problem of finding the threshold power whenever a new device is being
designed. Endeavor to build such an equation can frequently be found in the literature. For
example in [25] the authors gave a formula to calculate worst case SAR which used antenna
current or incident H-field as the parameter. In [21] a formula to calculate threshold power
for low directivity antennas has been given which uses frequency, separation distance, and
free space BW as the parameter.
3
1.2
EXPOSURE STANDARDS
Maximum exposure to electromagnetic waves is defined in terms of Specific
Absorption Rate (SAR). It is the absorbed power in per unit mass of tissue. There are two
widely known standards, which is being followed by the RF exposure regulation bodies of
respective countries. Exposure regulatory body in the United States (FCC or Federal
Communications Commission) has adopted the IEEE standard C95.1-2005 which restricts
maximum electromagnetic exposure to be 1.6 W/Kg averaged over 1-g cube of tissue for a
period of 30 min [22]. The same standard has also been adopted by Canada, Bolivia and
South Korea. This is slightly a stricter requirement than the ICNIRP standard which restricts
the maximum exposure to be 2 W/Kg averaged over 10g cube of tissue for a period of 6 min
[26]. ICNIRP guidelines are followed in Europe, Japan, Australia and most other countries.
1.3
IDENTIFYING REGIONS
Spatial region surrounding an antenna can be subdivided into three zones where the
fields have different configurations. At the very vicinity of the antenna reactive fields
predominate and hence this region is called the Reactive near-field region. The region next to
it is called the Radiating near-field (Fresnel) region, where “radiation fields predominate and
the angular field distribution is dependent upon the distance from the antenna”. The far most
zone is called the Far-field (Fraunhofer) region where distance does not play a part shaping
the angular field distribution. When an antenna is placed at a distance from a user it is
important to know which region of the antenna the user might fall into, so that
approximations may be made. We have well developed formula in the far-field region
although field configurations are not known to a very good approximation in the reactive
4
near-field or radiating near-field regions. This is because when we are trying to probe fields
in these regions we are actually altering the field configurations itself. And this alteration is
different for different electrical properties, shapes and sizes of the object near it.
Now separating these regions is not a very easy task as different antenna
configurations will have the regions shaped different than the others. For electrically small
antennas, that can be fit in a radian sphere of /2radius the far-field is defined outside this
sphere. The region inside the sphere is considered as reactive near-field region. No radiating
near-field region is defined for these antennas. For antennas that are larger than a
wavelength, far-field is defined beyond the distance
2D 2
. The reactive near-field region is

defined at a distance equal to or less than 0.62 D 3 /  from the antenna. The region in
between is the radiating near-field region.
Many of the RF devices available today are neither as small to fit in a radian sphere,
nor as large as a wavelength. For those devices we do not have a well-developed formula.
Besides, in a commercially available device, the geometry is so complex that it is more often
difficult to identify which part of the device is actually radiating. As we do not know the
effective aperture of those antennas, we cannot separate their field regions.
1.4
SELECTION OF FREQUENCY
Frequency bands of some commonly used portable wireless devices are given in
Table 1.1 [27] [28].
5
Table 1.1. Frequency bands of commonly used communication devices
fmin (MHz)
fmax (MHz)
Technology
810
958
PDC1
824
894
IS-1362
870
921
TETRA3
890
896
CMRS4
896
902
Fixed Land Mobile
902
928
Location &
Monitoring Service,
Amateur Radio
890
960
GSM
896
940
iDEN5
1217.37
1237.83
GPS [29]
1429
1501
PDC
1565.17
1585.63
GPS [29]
1710
1880
GSM
1850
1990
GSM
1920
2170
UMTS6
2300
2400
WiBro7
2390
2410
WLAN
2400
2484
Bluetooth
3700
4200
FSS8
4200
4400
Aeronautical Radio
Navigation
4990
5010
WLAN
5650
5850
Amateur Radio
5850
5925
GSO9/ NGSO10/ FSS
5925
6425
FSS
1. PDC (Personal Digital Cellular) is a Japanese cellular system. It uses Time Division
Multiple Access technique [30]
6
2. IS-136 is a technology for Digital AMPS mobile phone systems [31]
3. TETRA (Terrestrial Trunked Radio) is a specification for walkie-talkie [32]
4. CMRS (Commercial Mobile Radio Service) refers to the service that connects a
wireless network to the public switched telephone network
5. iDEN (Integrated Digital Enhanced Network) is a mobile phone system developed by
Motorola and Nextel [33]
6. UMTS (Universal Mobile Telecommunication System) is a third generation (3G)
standard for mobile communication systems which is widely known as W-CDMA
(Wideband Code Division Multiple Access) [34].
7. WiBro (Wireless Broadband) is an internet service technology used in Korea. It
follows the mobile WiMAX standard [35].
8. Geostationary satellites that are used for cable television, radio, telephony, and data
communications in north America are called FSS (Fixed Service Satellite) [36]
9. GSO – Geostationary Satellite
10. NGSO – Non Geostationary Satellite
1.5
COMPUTATIONAL METHOD
All the electromagnetic simulation software available today employ either of the
following techniques or a variant [37].
1. Differential methods
2. Integral methods
In differential methods, Maxwell’s equations are discretized and the values of the
unknowns are determined over the entire computational domain. This discretization might be
7
of the differential or the integral form (example – FIM or Finite Integral Method) of the
Maxwell’s equations. FEM (Finite Element Method), FDTD (Finite Difference Time
Domain), TLM (Transmission Line Method) fall into this category. Among the others, HFSS
employs FEM; CST, XFDTD employ FDTD technique. This method is very efficient for
complex geometries but suffers in accuracy for thin wires or thin layers which the integral
method does not.
In integral methods, integral form of the Green’s function of the structure is
discretized. IE3D involves MoM (Method of Moments) technique which falls into this
category. Boundary is imposed by the method in this case. Since Green’s function involves
the solution of sources of the fields, this method solves currents on the conducting region
(metal or PEC) only. Thus, it suffers in accuracy for geometries involving heterogeneous
dielectrics (example – a human head).
As my geometries were comparatively complex and had heterogeneous structures, I
planned to use the differential method technique.
The differential methods are further classified into two types: time domain and
frequency domain techniques. In time domain technique each element in the matrix of
variables is a “real” time varying quantity. The whole computational domain is discretized
into voxels (cubes). Electric fields are defined at the cell edges and magnetic fields at middle
of the face. The matrix of unknowns is then computed in a single voxel. Unknowns in the
neighboring voxels are then computed from the result obtained with this voxel, thus
numerically propagating the fields (Marching on Time –MoT). Frequency domain results can
also be obtained from this method using Fourier transform of the time domain results.
8
In frequency domain technique the excitation is sinusoidal. Time derivatives in the
Maxwell’s equations are replaced with j . A huge matrix of variables over the entire
domain is then formed and unlike the time domain technique, elements of the matrix might
be a complex quantity. This matrix is then solved for the unknowns of a single frequency.
For very large geometries solution of a single matrix might require more resource to perform
than with a time domain technique. Also if we need solution at different frequencies the time
domain technique would be more advantageous.
In this work, to model the antennas in free space I used a frequency domain technique
called HFSS. But while calculating SAR in the phantom, the computational domain being
very large (600 mm × 400 mm × 370 mm), I used a time domain technique called XFDTD
version 6.5 [38].
1.6
PHANTOM GEOMETRY AND PROPERTY
The shape, the size and the dielectric property of the phantom used conform to the
latest draft of International Electrotechnical Commission (IEC) [23]. The phantom is
elliptical in shape (Fig. 1.1) and large (600 mm × 400 mm) enough not to have any impact (<
1%) on SAR in 1-g or 10g volume. It was shown that a larger phantom gives conservative
SAR especially at the lower frequencies [24]. The phantom shell resembles an open top
container, filled with liquids having dielectric properties of human head. The properties of
this liquid was obtained by exposing a plane wave in the frequency range 300 MHz – 3000
MHz to an infinite half-spaced layered phantom which was achieved using permutations of
9
Figure 1.1. Homogeneous elliptical phantom
different tissue compositions and thickness to give the maximum peak spatial SAR [39] . The
same was validated for near-field exposure by replacing the plane wave with the fields of a
dipole and a generic phone (quarter wave monopole on a box) and the layered phantom with
an MRI based anatomical human head model. The liquid material properties thus obtained
was found to give an overestimate of SAR. Another benefit of using a homogeneous
phantom is that it is valid for the entire user group irrespective of their head sizes and shapes.
The bottom of the phantom was flat to facilitate maximum coupling between the device and
the phantom and thus give a conservative estimation of SAR. It has also been reported that
SAR induced in a flat phantom is higher compared to a realistic human head model or
spherical model [40]. Although at the distances considered here (40-200 mm) some of these
devices are handheld terminals, a hand model was not considered in my simulations or
measurements, since according to [41] a scenario not having a hand model would give a
conservative estimation of SAR. Dielectric properties of the liquid at different frequencies
10
are listed in Table 1.2. Tissue properties at 3700 MHz are obtained by interpolation of the
data of 3500 MHz and 4000 MHz.
Table 1.2. Electrical properties of tissue simulating liquids for homogeneous phantom
Frequency (MHz) Dielectric Constant,  r
Conductivity, 
900
41.5
0.97
1900
40
1.4
2450
39.2
1.8
3700
37.7
3.12
6000
35.1
5.48
Onishi et al. showed that the thickness, the loss tangent, the distance and the relative
permittivity of the shell also have an effect on SAR which is more pronounced at higher
frequencies ( > 3.8 GHz) [42]. They also concluded that the effect of relative permittivity
dominates the others. They suggested that if the relative permittivity of the shell is 4 with
tolerance ±1, the effect at 5.2 GHz can be reduced from ±30% to ±5%. In my work, the shell
had a relative permittivity value of 3.7 and conductivity close to zero thus having an
insignificant effect on SAR.
1.7
ANTENNA TO PHANTOM DISTANCES
Previously my colleague studied the cases where the phantom was at or less than 25
mm distances from the antenna [27]. The focus was mostly on mobile phone devices. To
include other hand-held and body-mounted devices I looked into the larger distances. The
11
standardized largest distance for those types of devices is 200 mm [23]. I studied two other
distances in between namely 40 mm and 100 mm to find the SAR rationale with distance.
12
CHAPTER 2
SAR INDUCED BY DIPOLE ANTENNA
2.1
ANTENNAS STUDIED FROM DIPOLE CLASS
Invented by Heinrich Rudolph Hertz in around the year 1886, dipole antenna is
one of the simplest antennas from the theoretical point of view [43]. Yet this antenna
would give us insight on how Specific Absorption Rate (SAR) can be rationalized in
terms of antenna geometry, frequency, and distance offering minimal mathematical
complexity.
Dipole antennas can be found to be of different lengths in the literature. The
shortest of them in terms of electrical length are often called an infinitesimal dipole or a
Hertzian dipole [44]. By definition they are very thin (a << ) and very short (l << /50)
antennas, or by other words the current distribution in them can be considered to be
constant for all practical purposes.
The next larger dipole antennas have a generic name “Small Dipole”. Dipoles of
lengths /50 < l < /10 fall into this category. Current distribution in these antennas can
be approximated as triangular waves.
The third kind in this antenna class is called “Finite Length Dipole”. These types
of antennas have sinusoidal current distribution. In this category half-wavelength
13
dipoles (l = /2) found most application in practice because having 73 ohms of radiation
resistance it can easily be matched with a coaxial transmission line.
In my research, dipoles of length /15 were considered from the second kind and
dipoles of length /2 were considered from the third kind. The reason for not considering
a larger dipole is that, for those antennas the current distribution on the antennas, and as a
result peak spatial SAR, has several peaks. Thus those antennas would have less 1-g and
10g averaged SAR. And the reason for not considering a dipole less than /15 length is
that as the size decreases bandwidth becomes narrower and almost no practical device
can be found with less bandwidth than that of a /15 dipole. At 2450 MHz a /15 dipole
has a BW of 0.4% and a Bluetooth device of the same frequency has a BW of 3.4% [27].
No antenna from the infinitesimal dipole category (or, first kind) was considered for the
same reason. I selected only one antenna from each category since the directivity is
almost similar for antennas of the same category.
2.2
FREQUENCY OF OPERATION
I judiciously chose some discrete frequencies to cover a vast range of devices and
technologies as given in Table 1.1. The frequencies were 900 MHz, 1900 MHz, 2450
MHz, 3700 MHz, and 6000 MHz. I did not go below 900 MHz because in that case the
computational domain becomes very large and thus needs very powerful computers to
simulate them especially at 200 mm distances. I did not also go above 6000 MHz since
even at that frequency a /15 dipole is only 3.33 mm long and for shorter physical
14
lengths accuracy becomes an issue because the dipoles might become largely unbalance
in terms of feed.
Figure 2.1. Dipole antenna prototypes for measurement
2.3
DIPOLE ORIENTATIONS
The orientations of the dipoles were such that the dipole arms were aligned with
the major axis of the elliptical phantom as discussed in Art 1.6. The tissue properties are
15
given in Table 1.2. Distances of 40, 100, and 200 mm were considered. Distance was
measured from the antenna feed point to the tissue simulating liquid as shown in Fig 2.2.
Figure 2.2. Definition of distance for dipole antennas
2.4
SIMULATION SETUP
Simulations were performed using XFDTD version 6.5 [38]. Liao absorbing
boundary with 40 cells uniform padding in all directions was used. A uniform
discretization of 1.0 mm was used as the base cell size. The dipoles were excited with
sinusoidal waves of the given frequencies of Art 2.2. Final results were established with 40 dB convergence. All the SAR values were normalized to 1 W of antenna input power.
Dipoles had a radius of 1.8 mm for frequencies up to (including) 2450 MHz and 0.5 mm
for frequencies above that.
16
2.5
MEASUREMENT PROCEDURE
The dipole antenna prototypes for measurement are shown in Fig. 2.1. All the
antennas were fed using a 50 Ω semi-grid coaxial cable of approximately 25 cm in length
except for the 6 GHz dipole. Coaxial cable feeding the 6 GHz dipoles had a length of
approximately 12 cm. Bazooka balun were used to choke off the current at the outer
conductor of the coax. As the dipoles shorter than /2 have very high return loss at the
feed points of the antennas, and a suitable method could not be found to get a normalized
value of SAR to 1 W of radiated power, SAR were not measured for those antennas.
2.5.1
Phantom Properties
An ELI4 phantom from SPEAG [45] was used for measurement. This phantom
fully complies with the guidelines of the latest draft IEC62209 Part 2 [23]. The
dimension of the ellipse was 600 mm by 400 mm as mentioned in Art 1.6. It was filled
with 28 liters of liquids having the properties as mentioned in that standard and a cover
prevents the liquids from evaporating. Thus the liquid depth was 150 mm. The dielectric
properties of the liquids were assured to be within ±5 % of the target value by using a
HP85070B dielectric probe kit in combination with a vector network analyzer prior to
SAR measurements. Whenever deviations were found liquids were re-adjusted so that
their properties were within ±5 % of the target value. A software version DASY 4.5 was
used.
17
Figure 2.3. An ELI4 Phantom
2.5.2
SAR Measurements
SAR measurements were performed in Seibersdorf Laboratories, Austria [46]. An
automated SAR testing facility with a 3 axis linear probe positioning unit was used to
measure SAR for all half wavelength dipoles of frequencies 900 MHz, 1900 MHz, 2450
MHz, 3700 MHz, and 6000 MHz at 40 mm, 100 mm, and 200 mm distances. The applied
SAR measurements and post-processing procedures all correspond to the requirements of
IEC62209-1 [47], and IEC62209-2 [23] respectively. Spatial peak SAR averaged over 1g as well as 10-g of mass were determined. No matching measure was implemented at the
input of the antennas. Radiated as well as reflected power was continuously monitored
during the measurements and all the SAR values were normalized to 1 W of input power
to the antenna.
18
Figure 2.4. Test chamber with wooden phantom holder and 3-axis probe positioning unit
Figure 2.5. Antenna holder with antenna below phantom
2.5.3
Measurement Uncertainty
The following factors are the major contributors for measurement uncertainty
19

Complete SAR uncertainty: Budget regarding measurement and post
processing of SAR data yields ± 25% for frequencies ≥ 800 MHz.

Power measurements: Total uncertainty of component calibration (cables,
couplers, attenuators) as well as power heads and power meters is ± 10%
Treating, in a first approach, the contributors as statistically independent and normally
distributed the overall uncertainty for SAR measurement is expected to be in the range of
approximately ± 25 – 30 %.
2.5.4
Measurement Equipment
The following table lists the measurement and test equipment used:
Table 2.1. Measurement equipment
Device
Type/
version
8753D
SN
Manufacturer
3410A04463
8722C
3232A02956
85052D
3101A04830
Amplifier
AR50S1G4
27947
Amplifier
1623
Amplifier
PA.RF
830960
BLWA
1050-200
20T4G18
Hewlett
Packard
Hewlett
Packard
Hewlett
Packard
Amplifier
Research
Parzich GmbH
Signal Generator
2024
112246/004
Signal Generator
SMR 40
100199
Power Meter
EPM Series
Power Meter
GB40201856
Vector Network
Analyzer
Vector Network
Analyzer
Calibration Kit SMA
Amplifier
882485
22501
20
Helmut Bonn
GmbH
Amplifier
Research
Marconi
Instruments
Rohde &
Schwarz
Agilent
Frequency
Range
30 kHz - 3
GHz
50 MHz – 40
GHz
DC - 26.5
GHz
800 MHz 4.2 GHz
830 MHz –
960 MHz
100 MHz –
500 MHz
4.2 GHz – 18
GHz
9 KHz – 2.4
GHz
10 MHz – 40
GHz
Power Head A (fwd.)
Power Head B (rev)
Directional Coupler 1
dB (fwd.)
Directional Coupler 2
dB (rev)
Cable
Attenuator 10 dB
Attenuator 20 dB
Dielectric Probe Kit
SAR Probe
SAR data acquisition
unit
Oval flat phantom
Probe positioning unit
SAR measurement
software
SAR post-processing
software
Precision Conical
Dipole
Spectrum Analyzer
2.6
Power
Sensor 8481
A
Power
Sensor 8481
A
4226-20
MY41091468
Agilent
10 MHz – 18
GHz
MY41091468
Agilent
10 MHz – 18
GHz
3693
Nerda
4226-20
3692
Nerda
500 MHz –
18 GHz
500 MHz –
18 GHz
Sucoflex 102
R412710124
R414720
85070B
19459
A0226
9149
US33020335
EX3DV4
Dasy 3 mini
3562
347
Suhner
Radiall
Radiall
Hewlett
Packard
SPEAG
SPEAG
ELI 4.0
3 axis linear
pos. system
SARAMS
1.5
SARPPS 1.2
1003+
01/2000
SPEAG
ARC-sr
PCD 8250
301/01
ARC-sr
E4450B
US40520766
Agilent
DC - 3 GHz
DC - 3 GHz
200 MHz –
20 GHz
ARC-sr
ARC-sr
80 MHz-2.4
GHz
9 kHz – 13.2
GHz
RESULTS
Table 2.2 summarizes computed and measured 1-g and 10-g averaged SAR data
for /15 and /2 dipoles placed at d=40, 100, and 200 mm from the phantom. These data
are normalized to 1W of power.
21
Table 2.2. Computed and measured SAR for  and  dipoles
Distanc
e, d
(mm)
Frequenc
y (MHz)
900
1900
2450
3700
6000
900
1900
2450
3700
6000
900
1900
2450
3700
6000
40
100
200
SAR (W/Kg) .
5.0
/15 dipole
Simulate Simulate
d SAR
d SAR
1g
10g
(W/Kg)
(W/Kg)
4.1849
2.9071
2.3867
1.5170
2.2925
1.3651
3.4847
1.7689
3.7859
1.6892
0.2204
0.1658
0.3766
0.2587
0.3483
0.2232
0.5044
0.2745
0.6096
0.2900
0.0642
0.0490
0.0751
0.0525
0.0956
0.0622
0.1184
0.0652
0.1511
0.0724
Simulate
d SAR
1g
(W/Kg)
2.2993
1.8655
1.9838
3.2823
3.9591
0.1978
0.3851
0.3497
0.5463
0.6775
0.0731
0.0798
0.0949
0.1289
0.1689
/2 dipole
Measure Simulate
d SAR
d SAR
1g
10g
(W/Kg) (W/Kg)
2.0730
1.6766
1.8560
1.2157
1.8400
1.1999
3.5100
1.7290
3.5100
1.7625
0.2030
0.1495
0.4180
0.2646
0.3740
0.2245
0.4500
0.2970
0.7520
0.3217
0.0820
0.0558
0.0780
0.0556
0.0970
0.0633
0.1150
0.0708
0.1730
0.0810
Measure
d
SAR 1g
(W/Kg)
1.5670
1.1260
1.1700
1.8400
1.4500
0.1530
0.2680
0.2590
0.2900
0.3020
0.0660
0.0530
0.0700
0.0810
0.0780
SAR vs Frequency - Dipole, d=40 mm
λ/15 (1g, sim)
λ/2 (1g, sim)
4.0
λ/2 (1g, meas)
3.0
λ/15 (10g, sim)
2.0
λ/2 (10g, sim)
λ/2 (10g,
meas)
1.0
0.0
500
1500
2500
3500 4500
Frequency (MHz)
5500
6500
Figure 2.6. SAR vs. frequency due to dipoles at d=40 mm distance from the
phantom
22
SAR (W/Kg) .
SAR vs Frequency - Dipole, d=100 mm
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
500
λ/15 (1g, sim)
λ/2 (1g, sim)
λ/2 (1g, meas)
λ/15 (10g, sim)
λ/2 (10g, sim)
λ/2 (10g, meas)
1500
2500
3500
4500
Frequency (MHz)
5500
6500
Figure 2.7. SAR vs. frequency due to dipoles at d=100 mm distance from the
phantom
SAR vs Frequency - Dipole, d=200 mm
SAR (W/Kg) .
0.20
λ/15 (1g, sim)
λ/2 (1g, sim)
0.15
λ/2 (1g, meas)
λ/15 (10g, sim)
0.10
λ/2 (10g, sim)
0.05
λ/2 (10g, meas)
0.00
0
1000
2000
3000 4000 5000
Frequency (MHz)
6000
7000
Figure 2.8. SAR vs. frequency due to dipoles at d=200 mm distance from the
phantom
Peak 1-g and 10-g averaged SAR for /15 and /2 dipoles at d = 40 mm are
shown in Fig. 2.6. Measurement and simulation results are in very good agreement. SAR
decreases sharply with frequency when the frequency is lower (below 2450 MHz). This is
23
because at these frequencies the 40 mm distance is still in the near-field region of the
antenna. Also because of this the SAR induced by /15 dipoles are higher than that
induced by /2 dipoles, particularly at the lower frequencies. This corroborates the
observations noted in [27]. Shorter dipoles have narrower bandwidths or larger quality
factors (Q) defined as [48]
Q  2
Maximum stored energy
Total energy lost period
Electrically smaller antennas (/15 dipoles) have more stored energy than the
electrically longer antennas (/2 dipoles), which results in higher SAR. But as frequency
increases SAR induced by /15 dipoles are almost the same as that by /2 dipoles. This
occurs because with increasing frequency the fixed physical distance becomes electrically
longer hence the phantom moves from the reactive near-field of the antenna to its
radiating near-field or the far-field. At higher frequencies a very small increase in SAR
can be observed for /2 dipoles compared to /15 which can be attributed to the former’s
slightly higher directivity (1.75 for /2 dipoles as opposed to 1.5 for /15 dipoles).
Furthermore, as the frequency increases beyond 2450 MHz, SAR also increases because
of increasing tissue conductivity at higher frequencies.
Figs. 2.7 and 2.8 show 1-g and 10-g averaged SAR induced by /2 and /15
dipoles at 100 and 200 mm distances. For frequencies higher than 2450 MHz, having
slightly higher directivities /2 dipoles induce slightly higher SAR than /15 dipoles. At
24
2450 MHz and d=100 mm, SAR is lower than that at adjacent frequencies. This is
probably because of the standing wave losses as the distance d=100 mm is close to a
wavelength at this frequency.
At 200 mm distance, the phantom is in the far-field of the antennas even at 900
MHz. Considering the r = /2formula for the field boundaries of electrically small
antennas the far-field boundary is 53 mm while using the 2D2/formula for electrically
large antennas this distance is 133 mm. Thus being in the far-field and having a slightly
higher directivity, the /2 dipoles always induce slightly higher SAR than the /15
dipoles. Fig. 2.9 illustrates the tissue conductivity as a function of frequency. Comparing
Figs. 2.8 and 2.9 we can clearly see how SAR in the far-field is influenced by tissue
conductivity.
6
Conductivity (S/m)
5
4
3
2
1
0
0
1000
2000
3000
4000
5000
Frequency (MHz)
Figure 2.9. Tissue conductivity vs. frequency plot
25
6000
7000
2.7
CONCLUSION
Spatial peak SAR averaged over 1-g and 10-g masses induced by /15 and /2
dipoles were computed using XFDTD. In the near-field region SAR data induced by the
shorter antennas are significantly higher than that induced by the longer antennas which
are in agreement with the findings reported in [27]. For phantoms in the far-field of
antennas the SAR induced by the longer antennas are slightly higher than that induced by
the shorter antennas. This is due to the slightly higher directivity of the longer antennas.
26
CHAPTER 3
SAR INDUCED BY MONOPOLE AND PLANAR INVERTED-F
ANTENNA (PIFA)
3.1
MONOPOLE ANTENNA MODELS
Linear quarter wavelength long wire (wire radius=1.8 mm) monopole antennas
operating at 900 MHz, 1900 MHz, and 2450 MHz were considered. Each monopole was
mounted on top of a metal box measuring 100 mm × 40 mm × 19 mm. The antenna was
placed at the center of the top surface of the metal box. The feed was connecting the
antenna and the metal box.
Figure 3.1. Monopole antenna and box model
27
3.2
PLANAR INVERTED-F ANTENNA (PIFA) MODELS
The PIFA’s studied in this work fall into two categories: (1) PIFAs in the air, and
(2) surface mount PIFAs. At 900 MHz and 1900 MHz antenna heights of 6 mm were
maintained from the ground plane without any dielectric in between. The geometrical
schematic of the 900 and 1900 MHz PIFAs is shown in Fig 3.2. Their dimensions are
listed in Table 3.1.
Figure 3.2. Geometry of planar inverted-F antenna for 900 and 1900 MHz
Table 3.1. PIFA dimensions for 900 and 1900 MHz
Frequency
W
L (mm)
s (mm)
t (mm)
d (mm)
(MHz)
(mm)
900
31
40
2.5
1
6
1900
13.5
20
2.0
1
6
28
Figure 3.3 shows the PIFA model for 2450 MHz. Corresponding parameter values
are listed in Table 3.2. Unlike the 900 and 1900 MHz PIFAs the PIFA for 2450 MHz was
a surface mount PIFA printed on FR4 substrate and its height from the top of the metal
box was smaller (3 mm as opposed to 6 mm for the former).
Figure 3.3. A surface mounted planar inverted-F antenna (PIFA) at 2450 MHz
Table 3.2. PIFA dimensions at 2450 MHz
Frequency
W (mm)
(MHz)
2450
3.3
Ws
L (mm)
(mm)
1
3
16
Ls (mm) L2 (mm)
24
8
s (mm)
d (mm)
1
3
SIMULATION SETUP
The elliptical phantom as described before was used. Monopoles were placed
such that their major axes were parallel to the major axis of the ellipse (phantom). Each
29
PIFA was placed such that the major axis of the metal box containing the PIFA was
parallel to the major axis of the phantom ellipse.
Antenna to phantom distances of 40 mm, 100 mm, and 200 mm were considered.
Distances were measured from the feed of the antenna to the tissue simulating liquid in
the phantom. A uniform discretization of 1.0 mm was used as the base cell size. Graded
mesh was used in some places. For example, a 0.5 mm discretization was used near the
monopoles and PIFA’s operating at 900 MHz and 1900 MHz; while a 0.25 mm
discretization was used near the 2450 MHz PIFA.
3.4
RESULTS
Table 3.3 summarizes the computed peak 1-g and 10-g averaged SAR for the
monopole antennas. If we examine any one frequency we find out that as the distance of
the antenna from the phantom increases both 1-g and 10-g SAR decreases as expected.
Table 3.3. SAR induced by monopole antennas
Frequency (MHz)
900
1900
2450
Distance (mm)
40
100
200
40
100
200
40
100
200
SAR 1g (W/Kg)
2.1243
0.1911
0.0765
1.5743
0.1670
0.0403
1.2477
0.3020
0.0831
30
SAR 10g (W/Kg)
1.5670
0.1448
0.0571
1.0300
0.1113
0.0252
0.7638
0.1935
0.0538
SAR vs Frequency - Monopole, d=40 mm
SAR (W/Kg) .
2.50
2.00
1.50
1.00
monopole 1g, sim
λ/2 dipole 1g, sim
monopole 10g, sim
λ/2 dipole 10g, sim
0.50
0.00
500
1000
monopole 1g, meas
λ/2 dipole 1g, meas
monopole 10g, meas
λ/2 dipole 10g, meas
1500
2000
Frequency (MHz)
2500
Figure 3.4. SAR vs. frequency due to monopoles at d=40 mm distance from the phantom
SAR vs Frequency - Monopole, d=100 mm
SAR (W/Kg) .
0.7
monopole 1g, sim
λ/2 dipole 1g, sim
monopole 10g, sim
λ/2 dipole 10g, sim
0.6
0.5
monopole 1g, meas
λ/2 dipole 1g, meas
monopole 10g, meas
λ/2 dipole 10g, meas
0.4
0.3
0.2
0.1
0.0
800
1300
1800
Frequency (MHz)
2300
Figure 3.5. SAR vs. frequency due to monopoles at d=100 mm distance from the
phantom
31
SAR vs Frequency - Monopole, d=200 mm
SAR (W/Kg) .
0.16
monopole 1g, sim
λ/2 dipole 1g, sim
monopole 10g, sim
λ/2 dipole 10g, sim
0.14
0.12
0.10
0.08
monopole 1g, meas
λ/2 dipole 1g, meas
monopole 10g, meas
λ/2 dipole 10g, meas
0.06
0.04
0.02
0.00
800
1300
1800
Frequency (MHz)
2300
Figure 3.6. SAR vs. frequency due to monopoles at d=200 mm distance from the
phantom
The computed and measured peak spatial 1-g and 10-g averaged SAR induced by
quarter-wavelength monopole antennas are compared with those by half wavelength
dipole antennas at 40, 100, and 200 mm distances in Figs. 3.4 – 3.6. As apparent,
simulation data are in excellent agreement with the measurement data. For all three
distances the SAR due to monopoles on boxes are always smaller than the SAR due to
dipoles. For d=40 mm the SAR variation for monopole antennas with frequency is
monotonous (decreases almost linearly). For d=100 mm the SAR versus frequency
characteristics for the monopoles is similar to that of the dipoles except for the inflection
point at 1900 MHz. The situation is nearly similar for d=200 mm.
32
(a)
(d)
(b)
(e)
(c)
(f)
Figure 3.7. 1-g averaged SAR distribution by 1900 MHz monopoles at (a) 40 mm, (b)
100 mm, and (c) 200 mm; and by 2450 MHz monopoles at (d) 40 mm, (e) 100 mm, and
(f) 200 mm distances from the antennas
Figure 3.8. 1-g averaged SAR distribution by 1900 MHz dipoles at (a) 100 mm, (b) 200
mm distances from the antennas
33
At 100, and 200 mm distances monopole 1-g and 10-g SAR decrease till 1900
MHz and then increase with frequency. This does not occur for the dipoles because the
SAR distributions for monopoles are different than dipoles. There are multiple SAR hot
spots (see Fig. 3.7) for the monopoles while there is a distinct one hot spot for the dipoles
(Fig. 3.8).
The 1-g and the 10-g averaged SAR induced by the PIFAs are given in the
following Table. The antennas were facing away from the phantom for all cases.
Table 3.4. SAR induced by PIFAs
Frequency (MHz)
900
1900
2450
Distance (mm)
SAR 1g (W/Kg)
SAR 10g (W/Kg)
40
2.9280
2.1338
100
0.1758
0.1328
200
0.0582
0.0417
40
0.8608
0.5317
100
0.1946
0.1280
200
0.0494
0.0329
40
0.4503
0.2577
100
0.0384
0.0233
200
0.0207
0.0116
34
SAR vs Frequency - PIFA, d=40 mm
λ/2 dipole 1g, sim
λ/2 dipole 10g, sim
PIFA 1g, sim
PIFA 10g, sim
λ/15 dipole 1g, sim
λ/2 dipole 1g, meas
λ/2 dipole 10g, meas
PIFA 1g, meas
PIFA 10g, meas
λ/15 dipole 10g, sim
SAR (W/Kg) .
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
850
1050
1250
1450
1650
1850
Frequency (MHz)
2050
2250
2450
Figure 3.9. SAR vs. frequency due to PIFAs at d=40 mm distance from the phantom
SAR vs Frequency - PIFA, d=100 mm
SAR (W/Kg) .
0.6
0.5
0.4
PIFA 1g, sim
PIFA 1g, meas
λ/2 dipole 1g, sim
λ/2 dipole 1g, meas
PIFA 10g, sim
PIFA 10g, meas
λ/2 dipole 10g, sim
λ/2 dipole 10g, meas
0.3
0.2
0.1
0
850
1050
1250
1450
1650
1850
2050
2250
2450
Frequency (MHz)
Figure 3.10. SAR vs. frequency due to PIFAs at d=100 mm distance from the phantom
35
SAR vs Frequency - PIFA, d=200 mm
0.14
PIFA 1g, sim
λ/2 dipole 1g, sim
PIFA 10g, sim
λ/2 dipole 10g, sim
SAR (W/Kg)
.
0.12
0.10
PIFA 1g, meas
λ/2 dipole 1g, meas
PIFA 10g, meas
λ/2 dipole 10g, meas
0.08
0.06
0.04
0.02
0.00
750
950
1150
1350
1550
1750
1950
2150
2350
Frequency (MHz)
Figure 3.11. SAR vs. frequency due to PIFAs at d=200 mm distance from the phantom
Computed and measured results of SAR induced by PIFAs in the conventional
orientation are compared with dipoles in Figs. 3.9 – 3.11. At d=40 mm the SAR due to
PIFAs decrease with frequency monotonously like that for monopoles on boxes. In
almost all cases the SAR due to PIFAs are smaller than the SAR due to /2 dipoles
except at 900 MHz, where SAR due to PIFAs are still lower than that due to /15 dipoles.
At 2450 MHz and d=200 mm, we see completely different trend in SAR induced by
PIFAs as compared to the /2 dipoles. The likely cause for this is the PIFAs at 2450
MHz are mounted on a FR4 substrate unlike the other PIFAs and thus the dielectric loss
in the substrate results in lower SAR. For example at d=200 mm and 2450 MHz, with 1
W of scaled net input power, power dissipated in tissue by a /2 dipole is 0.112 W. But
for the PIFAs half of the input power is lost in FR4 (0.504 W) and only 0.0283 W is
dissipated in tissue.
36
Figs. 3.12 – 3.14 compare the SAR due to PIFAs in the conventional and flipped
orientations at 40, 100, and 200 mm distances respectively. SAR is always higher in the
flipped orientation. For 900 MHz and at d=40 mm distance, SAR is higher in the
conventional orientation. The likely cause could be a near-field effect which is only
observed at 900 MHz. This was also the case with the 900 MHz PIFA at 20 mm as
reported in [27]. For the remaining cases the fact that PIFA in flipped orientation induces
higher SAR can be explained in terms of their physical distance from the phantom. In the
flipped orientation the nearest metal has a smaller distance from the phantom than in the
conventional orientation because they are mounted on a 10 mm thick metal box.
In general, it can be seen that for a constant antenna-to-phantom distance, SAR
decreases with frequency in the far-field region. This is because with increasing
frequency the antenna is actually placed at a larger electrical distance, i.e. the results
reflect the transition from near-field to far-field exposure. However, the trend of
monotonically decreasing SAR is not observable at 100 mm distance (see Fig. 3.13)
where higher SAR at 1900 MHz was obtained compared to the 900 MHz case. This
increase in SAR may be attributed to the standing wave effect at this distance (because
the distance is approximately close to half wavelength at 1900 MHz).
37
Figure 3.12. SAR vs. frequency of flipped PIFAs at 40 mm distance from the phantom
Figure 3.13. SAR vs. frequency of flipped PIFAs at 100 mm distance from the phantom
38
Figure 3.14. SAR vs. frequency of flipped PIFAs at 200 mm distance from the phantom
3.5
CONCLUSION
In this chapter, 1-g and 10-g averaged SAR induced by quarter wavelength
monopole antennas are compared with those of half wavelength dipoles. At lower
frequencies and closer distances, they have similar trends. But as the electrical or
physical distance increase, having smaller geometrical sizes dipoles exhibit far-field
phenomenon but the monopoles mounted on a large sized box do not. Thus we see some
difference in SAR trends induced by them. Besides as the monopoles do not have as
defined a radiation pattern as the dipoles, we see distributed SAR with monopoles.
PIFAs on the other hand show comparable SAR profile with dipoles except at
2450 MHz. This is because at 2450 MHz, a surface mounted PIFA was used unlike the
others. Apart from one data point (smallest distance and frequency), the SAR induced by
39
dipoles are always higher than those by conventionally oriented PIFAs as the former are
in closer distances of the phantom.
When conventionally oriented PIFAs are compared with their flipped
counterparts, we found the flipped ones induced higher SAR because the antenna to
phantom distance became closer, as the distance is defined from the antenna feed point to
the tissue liquid and the PIFAs are mounted on a 10 mm thick metal box.
40
CHAPTER 4
SAR INDUCED BY PATCH ANTENNA, DUAL BAND PIFA, AND
DUAL BAND IFA
4.1
MICROSTRIP PATCH ANTENNA MODELS
Each patch antenna was placed on a 100 mm × 40 mm × 10 mm metal box as
shown in Fig. 4.1. The patch was printed on an FR4 substrate. Antennas were designed
for operation in free space at the appropriate frequencies using HFSS. Antenna
dimensions are listed in Table 4.1. Patches were resonant at 1900, 2450, 3700, and 6000
MHz.
Figure 4.1. Schematic of a microstrip patch antenna on a metal box
41
Table 4.1. HFSS design data of microstrip patch antennas on FR4 substrate
Frequency
(MHz)
1900
2450
3700
6000
4.2
W (mm)
L (mm)
Ws (mm)
Ls (mm)
y (mm)
d (mm)
36.5
36.5
24.0
15.0
35.5
27.5
17.5
11.0
40
40
30
20
39
31
24
14
9.25
6.25
4.00
1.00
3
3
3
3
RESULTS (MICROSTRIP PATCH ANTENNA)
Figs. 4.2 – 4.4 summarize the computed peak 1-g and 10-g averaged SAR for
patch antennas in the conventional orientation.
SAR vs. Frequency - Patch conventional, d=40 mm
0.45
SAR1g, sim
SAR1g, meas
SAR10g, sim
SAR10g, meas
SAR (W/Kg)
.
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
0
1000
2000
3000
4000
Frequency [MHz]
5000
6000
7000
Figure 4.2. SAR vs. frequency of patch antennas at 40 mm distance from the phantom
42
SAR vs. Frequency - Patch conventional, d=100 mm
SAR (W/kg) .
0.09
SAR1g, sim
SAR1g, meas
SAR10g, sim
SAR10g, meas
0.08
0.07
0.06
0.05
0.04
0.03
0.02
0.01
0.00
0
1000
2000
3000
4000
Frequency (MHz)
5000
6000
7000
Figure 4.3. SAR vs. frequency of patch antennas at 100 mm distance from the phantom
SAR vs. Frequency - Patch conventional, d=200 mm
0.03
SAR (W/Kg) .
0.02
0.02
0.01
0.01
SAR1g, sim
SAR10g, sim
0.00
0
1000
2000
SAR1g, meas
SAR10g, meas
3000
4000
Frequency (MHz)
5000
6000
7000
Figure 4.4. SAR vs. frequency of patch antennas at 200 mm distance from the phantom
43
Comparing the SAR data shown in Figs. 4.2-4.4 with those shown in Figs. 2.6-2.8
it is clear that for a fixed distance and frequency the SAR induced by the microstrip patch
in the conventional orientation is significantly lower than that induced by dipole
antennas. The observation is similar if we compare Figs. 4.2-4.4 with the SAR induced
by monopoles and PIFAs as well. This is expected because the patch being a directional
antenna has its back lobe towards the phantom when radiating in the conventional
orientation.
1900 MHz
2450 MHz
3700 MHz
6000 MHz
Figure 4.5. SAR distribution of patch antennas in conventional orientation at 200
mm distance
1-g spatial averaged SAR distribution in a phantom is shown in Fig. 4.5 with
patch antennas at 200 mm distance from the phantom. We observe that at 6000 MHz
SAR is more distributed than at other frequencies and thus giving rise to a less peak
averaged value.
SAR results with flipped patches (radiating towards phantom) are shown in Figs.
4.6 – 4.8.
44
SAR vs. Frequency - Patch flipped, d=40 mm
6
SAR1g, sim
SAR1g, meas
SAR10g, sim
SAR10g, meas
SAR (W/Kg) .
5
4
3
2
1
0
0
1000
2000
3000
4000
Frequency (MHz)
5000
6000
7000
Figure 4.6. SAR vs. frequency due to flipped patch antennas at 40 mm distance
from the phantom
SAR vs. Frequency - Patch flipped, d=100 mm
1.20
SAR1g, sim
SAR1g, meas
SAR10g, sim
SAR10g, meas
SAR (W/kg) .
1.00
0.80
0.60
0.40
0.20
0.00
0
1000
2000
3000
4000
Frequency (MHz)
5000
6000
Figure 4.7. SAR vs. frequency due to flipped patch antennas at 100 mm
distance from the phantom
45
7000
SAR vs. Frequency - Patch flipped, d=200 mm
SAR (W/Kg) .
0.50
0.45
SAR1g, computed
SAR1g, measured
SAR10g, comuputed
SAR10g, measured
0.40
0.35
0.30
0.25
0.20
0.15
0.10
0.05
0.00
0
1000
2000
3000
4000
Frequency (MHz)
5000
6000
7000
Figure 4.8. SAR vs. frequency due to flipped patch antennas at 200 mm distance
from the phantom
In the flipped orientation SAR is always higher than that in the conventional
orientation as expected because the patch being highly directive higher absorption results
when it faces the phantom. If we use the term SAR enhancement factor to define the ratio
of the SAR in the flipped orientation to the SAR in the conventional orientation we can
clearly see that enhancement is higher at higher frequencies and longer distances. This is
because the antennas in those cases are clearly in the far-field and hence the directivity
comes into full play. For instance, the enhancement factor for 6 GHz and at 100 mm is
14.3 while that for 6 GHz and at 200 mm is 23.6. Similarly, the enhancement factor for
2.45 GHz at 100 mm distance is 10.3.
46
4.3
MODELS OF OTHER PLANAR ANTENNAS
A dual band Planar Inverted-F Antenna (PIFA) for operation at 2450/6000 MHz
bands was designed in HFSS. The schematic is shown in Fig. 4.9. This PIFA was printed
Figure 4.9. Schematic of a dual-band PIFA for operation in the 2450/6000 MHz bands
on a 3 mm thick FR4 (εr = 4.4, tan δ = 0.02) substrate. While computing SAR in XFDTD
an adaptive mesh of 0.25 mm was used near the antenna and 1.0 mm otherwise.
The schematic diagram of a dual-band Inverted-F Antenna (IFA) is presented in
Fig. 4.10. This antenna works at 2450 and 6000 MHz. The IFA is printed on a 3 mm
47
thick FR4 (εr = 4.4, tan δ = 0.02) substrate. In XFDTD an adaptive mesh of 0.25 mm was
implemented around the antenna.
Figure 4.10. A dual-band IFA for operation in the 2450/6000 MHz bands
4.4
SAR RESULT WITH OTHER PLANAR ANTENNAS
Computed peak 1-g and 10-g averaged SAR for mono band IFAs, dual band
PIFAs, and dual band IFAs are compared with those of patch antennas in Table 4.2 for
comparison at 2450 and 6000 MHz. Simulated data are marked as “sim”, experimentally
measured data as “exp”, and dual band as “DB”. At 2450 MHz, of all the antennas in
48
conventional orientation, Patch induces the lowest SAR. This is also true at most other
frequencies. This is because the focused beam of the patch antenna is directed away from
the phantom. At 6000 MHz however a dual band 2450/6000 MHz PIFA induces lower
SAR than a patch antenna. This is because the SAR induced by PIFA is more distributed
than the patch antenna.
Table 4.2. SAR Comparison of Dual Band PIFAs, Dual Band IFAs, Mon band IFAs with
Patch Antennas (Conventional Orientation)
Frequency
2450 MHz
Distance (mm)
40
DB PIFA SAR 1g, sim
0.421
DB PIFA SAR 1g, exp
0.572
DB IFA SAR 1g, sim
0.717
DB IFA SAR 1g, exp
1.089
IFA SAR 1g, sim
0.574
Patch SAR 1g, sim
0.244
Patch SAR 1g, exp
0.201
DB PIFA SAR 10g,
sim
0.240
DB PIFA SAR 10g,
exp
0.335
DB IFA SAR 10g, sim
0.414
DB IFA SAR 10g, exp
0.661
IFA SAR 10g, sim
0.333
Patch SAR 10g, sim
0.138
Patch SAR 10g, exp
0.119
100
0.063
0.119
0.151
0.271
0.130
0.042
0.039
200
0.019
0.035
0.058
0.070
0.051
0.014
0.015
0.038
0.010
0.075
0.091
0.177
0.078
0.025
0.022
0.023
0.030
0.047
0.024
0.009
0.010
6000 MHz
40
0.123
0.100
0.774
1.321
0.329
0.408
100
0.028
0.022
0.162
0.206
0.069
0.085
200
0.008
0.048
0.041
0.019
-
0.051
0.050
0.012
0.014
0.006
-
0.307
0.551
0.132
0.165
0.069
0.091
0.029
0.022
0.020
0.030
0.008
-
Peak spatial 1-g averaged SAR distribution for the dual band PIFA at 6000 MHz
in conventional orientation is shown in Fig. 4.11.
49
Figure 4.11. Peak Spatial 1-g averaged SAR distribution of dual
band PIFA at 6000 MHz at 40, 100, and 200 mm distances in
conventional orientation.
IFAs of 2450 MHz frequency are very common in laptop PCMCIA cards [49].
The IFA was placed in the conventional orientation with respect to the phantom and the
resulting SAR is listed. As the physical distance of the antenna from the phantom
increases, SAR decreases as expected.
Results in flipped orientation for patch antennas and dual band PIFAs are listed in
Table 4.3 at 2450 and 6000 MHz. The antennas induced higher SAR in the flipped
orientation when compared with the SAR in the conventional orientation for the same
antennas. This is because when the directive part of the antenna faces the phantom higher
absorption takes place. Also when flipped the nearest metal is closer to the phantom than
in the conventional orientation, as the distance is defined from the antenna feed to the
tissue simulating liquid.
50
Table 4.3. Comparison of SAR between Dual Band PIFAs and Patch Antennas (Flipped
Orientation)
Frequency
Distance (mm)
DB PIFA SAR 1g, sim
DB PIFA SAR 1g, exp
Patch SAR 1g, sim
Patch SAR 1g, exp
DB PIFA SAR 10g, sim
DB PIFA SAR 10g, exp
Patch SAR 10g, sim
Patch SAR 10g, exp
4.5
2450 MHz
6000 MHz
40
100
200
40
100
200
0.806
0.152
0.049
2.706
0.576
0.144
0.998
0.202
0.058
3.020
0.724
0.189
2.735
0.428
0.109
4.638
0.991
0.439
2.858
0.398
0.111
3.192
0.621
0.298
0.469
0.092
0.030
1.803
0.248
0.063
0.597
0.126
0.038
1.450
0.365
0.095
1.569
0.259
0.067
1.844
0.426
0.181
1.667
0.243
0.069
1.376
0.309
0.132
CONCLUSION
In this chapter a comparison between the SAR induced by patch antennas and
other planar antennas is given. It was found that patch antennas in the conventional
orientation induced the lowest SAR in most cases with some exceptions. The exceptions
can be explained in terms of the SAR distribution caused by these antennas. In the flipped
orientation, however, patch antennas were found to induce the highest SAR in all cases.
This is because, a patch antenna has a relatively focused beam, which when directed
toward the phantom results in higher SAR.
51
CHAPTER 5
SAR INDUCED BY ANTENNA ARRAYS
5.1
MODELS OF PATCH ANTENNA ARRAY
All antennas were designed for operation in free space using HFSS. Each patch
array was printed on a sufficiently large dielectric substrate with ground plane underneath
it. At or below 2450 MHz, FR4 was used as substrate. At 6000 MHz, FR4 being too lossy
RO4003 (εr = 3.38, tan δ = 0.0027) was used. Photographs of the prototypes are shown in
Fig. 5.1. Table 5.1 lists the peak realized gain of the patch arrays at different frequencies.
Figure 5.1. Patch array prototypes -- (a) on FR4 substrate, (b) on RO4003 substrate
52
Table 5.1. Peak Realized Gain of Different Patch Arrays
1900 MHz 2450 MHz 3700 MHz 6000 MHz
Peak Realized Gain (dB)
7.9
9.3
7.8
11.0
Peak Directivity (dB)
10.9
13.2
11.6
11.4
The patch array was placed in such a way that the plane of the patch or the ground
plane was parallel to the frontal plane of phantom ellipse. In the conventional orientation
the antenna faces away from the phantom and in the flipped orientation the antenna faces
the phantom. Antenna to phantom distances of 40, 100 and 200 mm were considered.
Distances were measured from the feed of the antenna to the tissue simulating liquid.
5.2
5.2.1
SAR RESULTS WITH PATCH ANTENNA ARRAY
Conventional Orientation
Figs. 5.2-5.4 summarize the computed peak 1-g and 10-g averaged SAR for patch
arrays in the conventional orientation.
SAR (W/Kg)
.
1.000
SAR 1g
SAR 10g
0.100
0.010
0.001
0.000
1500
2500
3500
4500
Frequency (MHz)
5500
6500
Figure 5.2. SAR vs. frequency of patch antenna arrays at 40 mm distance (conventional)
53
SAR (W/Kg) .
1.000
SAR 1g
0.100
0.010
0.001
0.000
1500
2500
3500
4500
Frequency (MHz)
5500
6500
Figure 5.3. SAR vs. frequency of patch antenna arrays at 100 mm distance (conventional)
SAR (W/Kg) .
0.010
0.001
SAR 1g
SAR 10g
0.000
1500
2500
3500
4500
5500
6500
Frequency (MHz)
Figure 5.4. SAR vs. frequency of patch antenna arrays at 200 mm distance (conventional)
For the 40 and 100 mm distances the phantom is essentially in the near-field of
the 1900 – 3700 MHz patch arrays. Therefore, the SAR shows a rapid increase with
frequency which is due to the increased tissue conductivity at higher frequencies.
Comparing 100 mm and 200 mm distances, it is clear that SAR plateaus at 2450 MHz for
200 mm than at 3700 MHz for 100 mm. Thus 200 mm makes the phantom to be more
into the radiating near-field or far-field of the array. The phantom is in the far-field at
6000 MHz for all distances.
54
5.2.2
Flipped Orientation
Computed and measured peak 1-g and 10-g SAR results for the patch antennas in
the flipped orientation are plotted in Figs. 5.5 – 5.7.
Figure 5.5. SAR vs. frequency of patch antenna arrays at 40 mm distance (flipped
orientation)
55
Figure 5.6. SAR vs. frequency of patch antenna arrays at 100 mm distance (flipped
orientation)
Figure 5.7. SAR vs. frequency of patch antenna arrays at 200 mm distance (flipped
orientation).
56
We observe relatively high SAR in the flipped orientation than in the
conventional orientation as expected. Between 3700 and 6000 MHz SAR increases
rapidly because the 6000 MHz patch has a less lossy substrate and the phantom is in the
far-field of the array. Also from 1900 MHz to 2450 MHz and at 40 mm distance SAR
decreases primarily because of a diffused distribution for the latter frequency as shown
below.
(b)
(a)
Figure 5.8. SAR distribution for flipped patch array at 40 mm distance for -- (a) 1900
MHz, (b) 2450 MHz frequencies
5.3
MODELS OF DIPOLE ANTENNA ARRAY
Dipole antenna arrays were designed for operation in free space using IE3D and
XFDTD. The 3-element and 5-element dipole arrays were simulated both with and
without reflectors, and the 7-element dipole array was simulated with a reflector.
57
5.3.1
A 3-element Dipole Array
IE3D Simulation Results in Free Space
Each dipole was made of a 10 mm wide metal strip and was fed at the center. The
dipole conductors consisted of perfect electric conductors. An arrangement of a 3element dipole array is shown in Fig. 5.9. This 3-element dipole array operating at 2450
MHz was simulated using IE3D. For operation against a reflector an infinite perfectly
conducting reflector was considered at distances of 20 and 40 mm from the array.
Simulation results for this dipole array are listed in
Table 5.2. Delta represents the
difference between the main lobe and the side lobe gain.
/2
20
25
10
51
/2
/2
Reflector
Figure 5.9. Schematic of a 3-element dipole antenna array
Table 5.2. IE3D simulation results for a 3-element dipole array
Reflector
spacing, h
(mm)
No reflector
20
40
G0 (dBi)
SLL (dBi)
Delta (dB)
5.49
11.48
8.57
-10
-3.5
-2
15.49
15
10.6
58
S11(dB)
-13.87
-16.09
-11.24
XFDTD Simulation with Finite Reflector
The schematic of a 3-element dipole array with a finite reflector is shown in Fig.
5.9. A finite reflector extending /2 distance from all sides of the dipole array was placed
at 20 mm distance from the array. This distance was chosen based on the observation in
Table 5.2.
Fig. 5.10 shows the gain pattern of the dominating E-field of the 3-element dipole array.
As can be observed from the figure,
Figure 5.10. Gain pattern of the dominating E-field of a 3-element dipole array
G0 = 12.17 dBi
SLL = -4.63 dBi
Back Lobe = -14.08 dBi
59
Radiation efficiency = 100%
System efficiency = 90.01%
So, D0 = 18.3 (linear)
The S-parameters were found as
  10.8  14.66  25.63
S     14.66  10.62  14.67 
 25.63  14.67  10.83 
XFDTD Simulation without Reflector
The above 3-element dipole array was simulated in XFDTD without a
reflector also. The resulting gain pattern is shown below.
Figure 5.11. Gain pattern of the dominating E-field of a 3 element dipole array without
reflector
60
G0 = 5.98 dBi
SLL = -7.46 dBi
Radiation efficiency = 100%
System efficiency = 93.51%
So, D0 = 4.24 (linear)
The S-parameters were found as
  19.73  18.02  26.89
S     17.97  20.56  17.97 
 26.89  18.08  19.73 
5.3.2
A 5-element Dipole Array
IE3D Simulation Results in Free Space
A 5-element dipole array at 2450 MHz frequency has the following gain
h (mm)
20
15
10
G0 (dBi)
13.37
13.69
13.13
SLL (dB)
0
0
-1
delta
13.37
13.69
14.13
S11 (dB)
-16.09
-10.41
-4.64
XFDTD Simulation with Finite Reflector
The setup of a 3-element dipole array, as shown in Fig. 5.9, was maintained
with two extra dipoles at the same 20 mm interval. The finite reflector was at 20 mm
distance from the antenna elements. Fig. 5.12 shows the gain pattern of the dominating Efield of the 5-element dipole array.
61
Figure 5.12. Gain pattern of the dominating E-field of a 5-dipole array with reflector
As can be observed from the figure,
G0 = 14.08 dBi
SLL = 0.17 dBi
Back Lobe = -12.57 dBi
Radiation efficiency = 100%
System efficiency = 90.8%
So, D0 = 28.18 (linear)
XFDTD Simulation with Finite Reflector
The above 5-element dipole array was simulated in XFDTD without a
reflector. The resulting gain pattern is shown Fig. 5.13.
62
Figure 5.13. Gain pattern of the dominating E-field of a 5-dipole array without reflector
G0 = 7.97 dBi
SLL = -4.93 dBi
Radiation efficiency = 100%
System efficiency = 92.49%
D0 = 6.77 (linear)
5.3.3
A 7-element Dipole Array
IE3D Simulation Results in Free Space
A 7-element dipole array against an infinite reflector at 20 mm distance from the
antenna elements was simulated in IE3D. At 2450 MHz frequency, G0 = 14.75 dBi, SLL
= 1 dBi.
63
XFDTD Simulation against Finite Reflector
Fig. 5.14 shows the gain pattern of the dominating E-field of the 7-element dipole array.
Figure 5.14. Gain pattern of dominating E-field of a 7-element dipole array with reflector
As can be observed from the figure,
G0 = 15.41 dBi
SLL = 1.97 dBi
Back Lobe = -11.47 dBi
Radiation efficiency = 100%
System efficiency = 89.46%
So, D0 = 38.85 (linear)
64
5.4
SAR RESULTS WITH DIPOLE ANTENNA ARRAY
Each dipole array was placed in such a way that the plane of the array or the
ground plane was parallel to the frontal plane of the phantom ellipse. The antenna arrays
were facing the phantom. Antenna to phantom distance of 200 mm was chosen. Table 5.3
summarizes the 1-g and 10-g SAR data obtained with the dipole arrays discussed above.
The said table also incorporates SAR results with patch antennas and patch antenna
arrays to find a rationale of SAR with directivity. Maximum dimension of the antenna
geometry is also listed in the Table.
Table 5.3. SAR induced by different antennas at 2450 MHz and at d=200 mm distance
Antenna Type
1 cylindrical /2 dipole
Single patch flipped
3-element /2 dipole
array, no reflector
5-element /2 dipole
array, no reflector
4-element patch array
3-element /2 dipole
array with reflector
5-element /2 dipole
array with reflector
7-element /2 dipole
array with reflector
D0
(linear)
SAR 1g
(W/Kg)
SAR 10g
(W/Kg)
1.75
2.28
0.0949
0.1088
0.0633
0.0677
Maximum
dimension
(mm)
51
108
4.24
0.1779
0.1087
196
6.77
0.108
0.0573
335
13.2
0.45
0.28
340
18.3
0.4816
0.2943
352
28.18
0.2529
0.1554
488
38.85
0.1699
0.1045
627
65
SAR vs D
0.6
SAR1g
SAR (W/Kg) .
0.5
SAR10g
0.4
0.3
0.2
0.1
0
0
5
10
15
20
25
30
35
40
45
Directivity, D0 (linear)
Figure 5.15. Scatter plot of SAR vs. Directivity
Fig. 5.15 shows the SAR scatter plot for all the antenna arrays considered in Table
5.3. The SAR is expected to increase with directivity if the phantom is in the far-field of
the antenna array. But we observe this trend only up to a directivity value of 18 with one
exception, namely the 5-element dipole array. We also see deviations from this trend
for highly directive antennas, 5-element and 7-element dipole arrays against
reflectors (D=28.18 and 38.85, respectively). The reason for this deviation is that the
effective aperture of these antennas are so large that even the 200 mm distance is
in the reactive near-field region of the antenna. It is intriguing to note that SAR is lower
when a highly directive antenna with a large effective aperture is about 200 mm from the
phantom than a smaller aperture directive antenna.
66
5.5
CONCLUSION
Peak 1-g and 10-g averaged SAR induced by microstrip patch arrays were
computed using XFDTD. Comparing the SAR of the arrays with the SAR of a single
patch in the conventional orientation we found that the SAR for the former was lower
than the latter. In the flipped orientation SAR is generally much higher. Discrepancies
were observed for several cases in the flipped orientation, when the SAR for a single
patch was found to be higher than that of the patch array (1900 MHz array at 40 mm,
2450 MHz patch array at 40 mm and 100 mm distances). For these cases, the
phantom being in the radiating near-field of the highly directive arrays altered the SAR
distributions resulting in more than a single hotspot. The SAR distribution is shown
in Fig. 5.8 for reference.
Later in this chapter, Peak 1-g and 10-g average SAR induced by dipole antenna
arrays were computed using XFDTD. Only one frequency 2450 MHz and one
distance 200 mm was chosen, and the effect of the directivity on SAR was
investigated. Generally for antennas having directivity less than 20, both 1-g and 10-g
SAR increases linearly with directivity. It appears that both the 1-g and 10-g SAR due to
a 5 element array (directivity = 6.77 dBi) are outliers. It can also be observed that
increasing directivity does not necessarily mean higher SAR. This is because the aperture
size is either comparable or larger than the phantom size.
67
CHAPTER 6
THRESHOLD POWER RATIONALE
6.1
THRESHOLD POWER RATIONALE AT 40 MM DISTANCE
In [21] the authors developed a rationale of threshold power with frequency,
distance, and bandwidth (BW). Among these parameters a rationale with the distance is
more straight-forward, since as the distance increases the phantom is less exposed to
electromagnetic exposure from the antenna. After that comes the effect of the frequency.
Frequency affects the SAR in two ways. First, as the frequency increases, electrical
distance of the phantom from the antenna increases. So SAR should have a tendency to
decrease. But on the other hand, tissue conductivity of the phantom increases with
frequency. So SAR should increase with frequency. Thus effect of frequency on SAR is
characterized by which of these two phenomena is dominant over the other. For the case
in [27], antenna to phantom distance was 25 mm maximum. At these distances the
phantom was in the near-field of the antennas in most cases. Now, the 7 dB bandwidth is
reciprocal of the Quality factor (Q) which is a function of the stored and the radiated
energy from the antenna. So it makes sense that SAR depends on BW in this case.
68
(a)
(b)
(c)
(d)
(e)
Figure 6.1. Threshold Power for 10-g SAR at 40 mm distance from the antenna vs.
antenna free space -7 dB BW at (a) 900 MHz, (b) 1900 MHz, (c) 2450 MHz, (d) 3700
MHz, (e) 6000 MHz
In this work we attempted to develop the rationale for threshold power for
distances larger than 40 mm. In Fig. 6.1, antenna threshold power data corresponding to
the 10-g SAR for a phantom at 40 mm distance from the antenna vs. the free space
69
antenna bandwidth are plotted. The threshold power was calculated using the following
formula [50]
Pth, m 
Pt SARlim it , m
SARm
As can be seen from these plots, the relationship is not linear as it was in [27]. A
power relationship provides a relatively better fit to the data but not good enough since
average R2=0.28. We concluded that at this distance a relationship between the threshold
power and bandwidth is not apparent. Table 6.1 lists the lowest 10-g threshold powers for
low directivity antennas in the 900–6000 MHz frequency range.
Table 6.1. Lowest Pth,10g at 40 mm distance
Frequency
(GHz)
0.9
1.9
2.45
3.7
6
Pth,10g
(mW)
687.9708
1318.392
1465.094
1130.646
1109.385
Antenna
/15 dipole
/15 dipole
/15 dipole
/15 dipole
Dual-Band PIFA Flipped
Except at 6 GHz, /15 dipoles induce the highest SAR at all frequencies or has
the lowest threshold power compared to the other antennas. The same phenomenon was
also observed in [21]. The reason for the 6 GHz case being an exception is that at 40 mm
distance, some of the antennas are no longer in the near-field and thus directivity comes
70
into effect. The dual band PIFA in the flipped orientation has a directivity of 3.42 (linear)
as compared to 1.46 (linear) for the /15 dipole at 6 GHz.
A scatter plot of the data in Table 6.1 is shown in Fig. 6.2. As can be seen, a third
order polynomial in frequency fits quite well with all of these data points (R2 = 0.9897).
1600
1400
Pth,10g (mW)
1200
1000
800
600
R² = 0.9897
400
0
2
4
6
Frequency (GHz)
Figure 6.2. Lowest Pth,10g vs. frequency at 40 mm distance
Upon suitable adjustment we get the following formula which underestimates the
threshold power for all the antennas studied in this thesis when the phantom is at a 40
mm distance from the antenna.
Pmax,10g  60.883 f 3  680.47 f 2  2209.3 f  986.66
(1)
The distance was defined from the antenna feed to the phantom liquid. In practical
cases the distance would most likely be measured from the nearest metal part of the
71
transmitting device to the phantom. Thus, this formula would give a more conservative
estimate of SAR for such a practical scenario.
6.2
THRESHOLD POWER RATIONALE AT 100 – 200 MM DISTANCES
At 100 – 200 mm distances threshold power becomes a function of directivity
(D), as well as of frequency and separation distance. As was argued in [21], to make the
formula more practical the same variable s has been used. This is the separation distance
between the phantom liquid and the nearest metal part of the antenna. And as has been
discussed, this only makes the formula more conservative. The relationships between
Pth,10g and the parameters f, s, and D are illustrated in Figs. 6.3 – 6.5. In Fig. 6.3, Pth,10g of
a /2 dipole antenna is plotted at 100 mm and 200 mm distances. We can see that the
variation of Pth,10g can be modeled as a third order polynomial in frequency. In Fig. 6.4,
Pth,10g of a /2 dipole antenna is plotted as function of separation distance s. Thus, Pth,10g
can be approximated as a linear function of s. Finally, in Fig. 6.5 variation of Pth,10g has
been shown in 900 MHz and at 100 mm and 200 mm distances as a function of antenna
free space directivity toward intended direction of operation. A second order polynomial
is used to model the data.
72
Thousands
40
d=100mm
d=200mm
35
30
Pth (mW)
25
20
15
10
5
0
0
1000
2000
3000
4000
Frequency (MHz)
5000
6000
7000
Thousands
Figure 6.3. Relationship of Pth,10g (mW) with frequency at d=100 and 200mm for /2
dipoles
40
f=900 MHz
f=1900 MHz
f=2450 MHz
f=3700 MHz
f=6000 MHz
35
30
Pth(mW)
25
20
15
10
5
0
0
50
100
150
Separation(mm)
200
250
Figure 6.4. Relationship of Pth,10g (mW) with separation distance, s
73
x 10000
6
d=200mm
d=100mm
5
Pth (mW)
4
3
2
1
0
1.3
1.4
1.5
1.6
Directivity (Linear)
1.7
1.8
1.9
Figure 6.5. Relationship of Pth,10g (mW) with directivity at d=100 and 200mm
We then performed suitable regression analysis considering these observations.
Finally we made some adjustments to develop a formula which will give a conservative
estimate of threshold power for compliance with regulatory standards. The formula is
given below.
Pmax,10g  3.56 10 4  1.8175 10 2 f 3  2.426 103 f 2  1.027110 4 f
 7.7654 103 D 2  3.7235 10 4 D  2.2465 10 2 S
(2)
where, f is the intended frequency of operation in GHz, D is the free space linear
directivity of the antenna toward the phantom, and s is separation distance. Pmax,10g then
gives the threshold power in mW.
74
6.3
UNDERESTIMATION
To compare the calculated threshold power and the radiated power found in this
work, underestimation is calculated in dB for all the antennas studied here. The same
underestimation formula as given in [21] has been used.
Underestimation  10 log10 ( Pth,10g / Pmax,10g )
The results are sub-divided into two sections. The first section discusses the
underestimation at 40 mm distance and the second section discusses the underestimation
at 100 – 200 mm distance.
6.3.1
Underestimation at 40 mm Distance
Fig. 6.6 gives a comparison between the threshold power Pth,10g and the power
calculated using equation (1). We can see that although the underestimation varies for
different antennas, all the values are positive. A positive underestimation means if the
antenna threshold power is equal to or less than Pmax,10g the antenna would be compliant
with the SAR limits. The maximum underestimation was found to be 16.3 dB and the
minimum was 0.5 dB. It can be seen that the lowest underestimation of Pth,10g is
consistent with frequency, so the formula is not biased toward a high or low frequency
zone.
75
18
Underestimation of Pth,10g (dB)
16
14
12
10
8
6
4
2
0
0
1
2
3
4
Frequency (GHz)
5
6
7
Figure 6.6. Underestimation using (1) at 40 mm distance
6.3.2
Underestimation at 100 – 200 mm Distance
Underestimation using formula (2) at 100 mm and 200 mm distances are plotted
in Fig. 6.7 and 6.8. This formula does not consider patch antennas in the flipped
orientation so their threshold power is not compared in these plots. The reason why the
patch antennas were excluded was because at 100–200 mm distances the effect of
directivity is more dominant. And since patch antennas in the flipped orientation have
higher directivities they have to be dealt separately from low directivity antennas. For
comparison, the 3700 MHz patch antenna in the flipped orientation had a directivity of
4.78 (linear) while the same for the /2 dipole was only 1.81 (linear). Therefore the
formula in (2) was developed considering low directivity antennas, such as dipoles,
monopoles, PIFAs, IFAs etc.
76
15
λ/2 dipole
Underestimation (dB)
λ/15 dipole
10
monopole
PIFA_conv
PIFA_fl
Patch_conv
5
dB_PIFA_Conv
dB_PIFA_fl
0
500
1500
2500
3500
4500
Frequency (MHz)
5500
6500
Figure 6.7. Underestimation using (2) at 100 mm distance
10
Underestimation (dB)
λ/2 dipole
λ/15 dipole
monopole
PIFA_conv
5
PIFA_fl
Patch_conv
dB_PIFA_conv
DB_PIFA_fl
0
500
1500
2500
3500
4500
Frequency (MHz)
5500
6500
Figure 6.8. Underestimation using (2) at 200 mm distance
As can be seen from the above figures, formula (2) always underestimates the
threshold power corresponding to the 10-g SAR. since all the underestimation values are
positive. Maximum and minimum underestimation obtained with this formula at 100 mm
77
distance is 11.87 dB and 0.7 dB respectively. At 200 mm distance they are 8.28 dB and
2.5E-5 dB respectively.
6.4
THRESHOLD POWER RATIONALE FOR DIRECTIVE ANTENNAS
Formulae (1) and (2) were developed for low directivity antennas, since
incorporation of the highly directive antennas results in higher rms error. However, a
number of directive antennas were also studied. To make the antennas highly directive
antenna arrays were used. As described in Chapter 5, simulations and measurements were
performed on a 4 element patch array in conventional and flipped orientations in the
frequency range of 1900–6000 MHz and at 40, 100, and 200 mm distances. SAR induced
by 3, 5, and 7 elements dipole arrays were also computed at 2450 MHz and 200 mm
distance to find the relationship between the induced SAR and the directivity for highly
directive antennas.
Fig. 6.9 shows the scatter plots of threshold power for compliance with 10-g peak
spatial averaged SAR limits. The data belong to the frequency of 2450 MHz and a
phantom 200 mm away from each antenna. We normally expect the SAR to increase with
directivity if the phantom is in the far-field of the antenna. But we observe this trend only
up to a directivity value of 18 or less. We see deviations for highly directive antennas,
namely 5-element and 7-element dipole arrays backed by reflectors (D=28.18 and 38.85
respectively). The reason for this deviation is that the effective aperture of these antennas
are so large that even the 200 mm distance is in the reactive near-field region of the
antenna. It is intriguing to note that SAR is lower when highly directive antennas with
large effective apertures are about 200 mm away from the phantom than a smaller
78
aperture and lower directivity antenna. Furthermore, if we look closely we see that the
threshold power can be approximated with a second order polynomial in free space
directivity. Had we had enough data at other distances and frequencies, a formula to
estimate threshold power similar to the previous ones could be developed.
35000
y = 53.884x2 - 2381.3x + 32406
R² = 0.8697
30000
Pth,10g (mW)
25000
20000
15000
10000
5000
0
0
5
10
15
20
25
30
35
40
Directivity (Linear)
Figure 6.9. Pth,10g vs. directivity of high directivity antennas
6.5
CONCLUSION
Methods to estimate maximum allowable threshold power levels for antennas at
distances of 40, 100, and 200 mm from phantoms were developed. Two formulas were
developed. One is applicable to devices that are 40 mm away from a user. The second
one is applicable to devices whose normal mode of operation is at 100 – 200 mm from
79
the user. In the latter formula a relationship of threshold power has been established with
antenna free space directivity along with other parameters.
80
CHAPTER 7
CONTRIBUTION AND FUTURE WORK
7.1
CONTRIBUTION
A large number of antennas, such as – dipoles, monopoles, PIFAs, IFAs, patches,
dual-band antennas, arrays etc., have been studied in this thesis that encompasses a large
variety of devices in the 900 – 6000 MHz frequency range. SAR was evaluated using the
FDTD method and was validated with experimental data. A large homogeneous elliptical
phantom was used (unlike the phantoms most commonly used in the literature) to ensure
a conservative exposure. This work addresses the devices that do not operate at the close
near-field distances of a user. Examples of such devices include vehicular electronics,
laptop PCMCIA cards, Bluetooth dongles, wireless LAN routers, cordless phone base
stations, pico - base stations, GPS devices etc. The contribution of this work lies in the
following findings:
It was found out that at close near-field distances antennas having smaller
geometries induce distinguishably higher SAR than the antennas having larger
geometries but as the user moves to the far-field only a small difference between the two
can be observed. This can clearly be seen from the SAR results of /15 and /2 dipoles.
81
Highly directive antennas when facing the phantom at far-field distances induce
higher SAR but as the phantom is taken to the near-field distance the same antennas
induce much lower SAR.
Based on these and other observations two formulas were developed to estimate
the threshold power that meet the SAR limits. Both of these formulas are relevant for low
directivity antennas and are applicable to 40 mm, and 100-200 mm distances.
Studies were also conducted to understand the SAR exposure from directive
antennas. It was observed that for antennas having directivity less than 20, both 1-g and
10-g SAR increased linearly with directivity. SAR did not increase when the directivity
was increased further primarily because the phantom moved into the near-field region of
the antenna, the aperture size of which became either comparable or larger than the
phantom size.
7.2
FUTURE WORKS
More distances, frequencies, and high directivity antennas need to be studied both
numerically and experimentally to develop a formula for highly directive antennas.
As the electrical distance between the device and the phantom increases it
becomes more difficult to clearly define the boundaries of the near and far-zones. Studies
need to be done to clearly define the near and far zone boundaries for high directivity
antennas, antenna arrays, and antennas that are in between the definitions of electrically
small and large antennas.
82
Temperature rise studies should be performed in true anatomical human models to
understand the maximum temperature rise and the temperature distribution due to
exposure from these and other antennas.
83
REFERENCES
[1] W. B. Deichmann, F. H. S. Jr, M. Keplinger, and K. F. Lampe, "Acute effects of
microwave radiation on experimental animals (24,000 megacycles)," J. of
Occupational Med., vol. 1, Jul. 1959.
[2] S. Michaelson, R. Thomson, and J. Howland, "Physiologic aspects of microwave
irradiation of mammals," Amer. J. Physiol., vol. 201, p. 351, 1961.
[3] W. B. Deichmann, J. Miale, and K. Landeen, "Effect of microwave radiation on the
hemopoietic systems of the rat," Toxicology and App. Pharmacology, vol. 6, pp.
71-77, 1964.
[4] G. M. Samaras, L. R. Muroff, and G. Anderson, "Prolongation of life during highintensity microwave exposures," IEEE Trans. Microwave Theory Tech., vol. 19, no.
2, pp. 245-247, Feb. 1968.
[5] T. L. Pay, "Microwave effects on reproductive capacity and genetic transmission in
drosophila melanogaster," J. Microw. Power, vol. 7, p. 175, 1972.
[6] A. W. Guy, J. C. Lin, P. O. Kramar, and A. F. Emery, "Effect of 2450 MHz
radiation on the rabbit eye," IEEE Trans. Microwave Theory Tech., vol. 23, pp.
492-498, 1975.
[7] J. C. Monahan and H. S. Ho, "The effect of ambient temperature on the reduction
of microwave energy absorption by mice," Radio Science, vol. 12, no. 6, pp. 257262, 1977.
[8] R. A. Tell and F. Harlen, "A review of selected biological effects and dosimetric
data useful for development of radiofrequency safety standards for human
exposure," J. Microw. Power, vol. 14, no. 4, pp. 405-424, 1979.
[9] J. O. deLorge, "Operant behavior and rectal temperature of squirrel monkeys during
2.45-GHz microwave radiation," Radio Science, vol. 14, pp. 217-225, 1979.
84
[10] M. I. Gage, "Microwave irradiation and ambient temperature interact to alter rat
behavior following overnight exposure," J. Microw. Power, vol. 14, no. 4, pp. 389398, 1979.
[11] M. Barbiroli, C. Carciofi, V. D. Esposti, and G. Falciasecca, "Evaluation of
Exposure Levels Generated by Cellular Systems: Methodology and Results," IEEE
Trans. on Veh. Tech., vol. 51, no. 6, Nov. 2002.
[12] F. Lacroux, A. C. Carrasco, A. Gati, M. F. Wong, and J. Wiart, "SAR and averaged
power density near a UMTS base-station antenna," in IEEE MTT-S Int. Mic. Sym.,
June 2006, pp. 1987-1990.
[13] K. R. Foster, "Radiofrequency Exposure From Wireless Lans Utilizing Wi-Fi
Technology," Health Physics, vol. 92, no. 3, pp. 280-289, Mar. 2007.
[14] Y. Zhou, J. Streckert, H. N. Mbonjo, and V. Hansen, "SAR of wireless
communication terminals opearated near the human body using the example of
PCMCIA data cards," in Progress in Electromag. Research Symp., Hangzhou,
China, 2008, pp. 745-749.
[15] P. Bernardi, M. Cavagnaro, S. Pisa, and E. Piuzzi, "Specific absorption rate and
temperature increases in the head of a cellular-phone user," IEEE Trans.
Microwave Theory Tech., vol. 48, no. 7, pp. 1118-1126, July 2000.
[16] C. W. Jang, "Portable communication device for minimizing specific absorption
rate (SAR) value of electromagnetic wave," Patent US 2002/0187806 A1, Dec. 12,
2002.
[17] W. G. Whittow and R. M. Edwards, "A study of changes to specific absorption
rates in the human eye close to perfectly conducting spectacles within the radio
frequency range 1.5 to 3.0 GHz," IEEE Trans. Antennas Propagat., vol. 52, no. 12,
pp. 3207-3212, Dec. 2004.
[18] A. Y. Simba, T. Hikage, S. Watanabe, and T. Nojima, "Specific absorption rates of
anatomically realistic human models exposed to RF electromagnetic fields from
mobile phones used in elevators," IEEE Trans. Microwave Theory Tech., vol. 57,
no. 5, pp. 1250-1259, May 2009.
[19] T. Hondou, "Rising level of public exposure to mobile phones: accumulation
through additivity and reflectivity," J. Phys. Soc. Jpn., vol. 71, no. 2, pp. 432-435,
Feb. 2002.
[20] A. Kramer, J. Frohlich, and N. Kuster, "Towards danger of mobile phones in
planes, trains, cars, and elevators," J. Phys. Soc. Jpn., vol. 71, no. 12, p. 3100, Dec.
85
2002.
[21] A. T. M. Sayem, M. G. Douglas, G. Schmid, B. Petric, and M. Ali, "Correlating
threshold power with free-space bandwidth for low directivity antennas," IEEE
Trans. Electromag. Comp., vol. 51, no. 1, pp. 25-37, Feb. 2009.
[22] "IEEE Standard for Safety Levels With Respect to Human Exposure to Radio
Frequency Electromagnetic Fields, 3 kHz to 300 GHz," IEEE Standard C95.12005, 2006.
[23] "Human exposure to radio frequency fields from handheld and body-mounted
wireless communication devices - Human models, Instrumentation, and Procedures
- Part 2: Procedure to determine the SAR for mobile devices in close proximity (30
MHz to 6 GHz)," IEC, Committee draft 62209-2, October 2008.
[24] A. K. Lee, H. D. Choi, B. C. Kim, H. S. Lee, and J. K. Pack, "Effect of head size
for mobile phone exposure on EM absorption," in APMC, Taipei, 2001, pp. 384387.
[25] N. Kuster and Quirino Balzano, "Energy absorption mechanism by biological
bodies in the near field of dipole antennas above 300 MHz," IEEE Trans. Veh.
Technol., vol. 41, no. 1, pp. 17-23, Feb 1992.
[26] ICNIRP, "Health issues related to the use of hand-held radiotelephones and base
transmitters," Health Physics, vol. 70, no. 4, pp. 587-593, April 1996.
[27] Abu T. M. Sayem, "Designing electrically small antennas and the effects of their
radiation on humans," University of South Carolina, Columbia, PhD Thesis 2007.
[28] "Spectrum Inventory Table, 137 MHz to 100 GHz," Federal Communications
Commission, Washington, DC, DA96-1704, 1996.
[29] "Navstar GPS space segment/ Navigation user interfaces," ARNIRC Engineering
Services, LLC, El Segundo, CA 90245, Interface Specification IS-GPS-200D, 7
Dec 2004.
[30] (2010, May) About Wireless. [Online]. HYPERLINK "http://www.aboutwireless.com/terms/pdc.htm" http://www.about-wireless.com/terms/pdc.htm
[31] Wikipedia contributors. (2010, May 7) Digital AMPS --- Wikipedia, The Free
Encyclopedia. [Online]. HYPERLINK
"http://en.wikipedia.org/w/index.php?title=Digital_AMPS&oldid=338213217"
http://en.wikipedia.org/w/index.php?title=Digital_AMPS&oldid=338213217
86
[32] Wikipedia contributors. (2010, May 7) Terrestrial Trunked Radio --- Wikipedia,
The Free Encyclopedia. [Online]. HYPERLINK
"http://en.wikipedia.org/w/index.php?title=Terrestrial_Trunked_Radio&oldid=358
977569"
http://en.wikipedia.org/w/index.php?title=Terrestrial_Trunked_Radio&oldid=3589
77569
[33] (2010, May 7) About Wireless. [Online]. HYPERLINK "http://www.aboutwireless.com/terms/iden.htm" http://www.about-wireless.com/terms/iden.htm
[34] Wikipedia contributors. (2010, May 7) W-CDMA (UMTS) --- Wikipedia, The Free
Encyclopedia. [Online]. HYPERLINK
"http://en.wikipedia.org/w/index.php?title=WCDMA_(UMTS)&oldid=359096096"
http://en.wikipedia.org/w/index.php?title=W-CDMA_(UMTS)&oldid=359096096
[35] Wikipedia contributors. (2010, May 7) WiBro -- Wikipedia, The Free
Encyclopedia. [Online]. HYPERLINK
"http://en.wikipedia.org/w/index.php?title=WiBro&oldid=357774691"
http://en.wikipedia.org/w/index.php?title=WiBro&oldid=357774691
[36] Wikipedia contributors. (2010, May 7) Fixed Service Satellite --- Wikipedia, The
Free Encyclopedia. [Online]. HYPERLINK
"http://en.wikipedia.org/w/index.php?title=Fixed_Service_Satellite&oldid=329697
035"
http://en.wikipedia.org/w/index.php?title=Fixed_Service_Satellite&oldid=3296970
35
[37] (2010, May 10) Advanced EM. [Online]. HYPERLINK
"http://www.advancedem.com/methods.html"
http://www.advancedem.com/methods.html
[38] (2010, May) Remcom. [Online]. HYPERLINK "http://www.remcom.com/"
http://www.remcom.com/
[39] A. Drossos, V. Santomaa, and N. Kuster, "The dependence of electromagnetic
energy absorption upon human head tissue composition in the frequency range of
300-3000 MHz," IEEE Trans. Microw. Theory Tech., vol. 48, no. 11, pp. 1988–
1995, Nov 2000.
[40] M. Okoniewski and M. A. Stuchly, "A study of the handset antenna and human
body interactin," IEEE Trans. Microw. Theory Tech., vol. 44, no. 10, pp. 1855–
1864, Oct. 1996.
87
[41] F. Gemperle, C. Kasabach, J. Stivoric, M. Bauer, and R. Martin. (2010, May)
Design for Wearability, Carnegie Mellon University, PA (USA). [Online].
HYPERLINK "http://www.ices.cmu.edu/designtimeline/wearability/files/Wearability.pdf" http://www.ices.cmu.edu/designtimeline/wearability/files/Wearability.pdf
[42] T. Onishi, and S. Uebayashi, "Influence of phantom shell on SAR measurement in
3-6 GHz frequency range," IEICE Trans. Commun., vol. E88-B, no. 8, pp. 32573262, August 2005.
[43] Wikipedia contributors. (2010 19:58 UTC, May 6) Dipole antenna --- Wikipedia,
The Free Encyclopedia. [Online]. HYPERLINK
"http://en.wikipedia.org/w/index.php?title=Dipole_antenna&oldid=356773637"
http://en.wikipedia.org/w/index.php?title=Dipole_antenna&oldid=356773637
[44] Constantine A. Balanis, Antenna Theory, 3rd ed. Hoboken, New Jersey, USA: John
Wiley & Sons, Inc., Publication, 2005.
[45] (2010, May) SPEAG, Schmid & Partner Engineering AG. [Online]. HYPERLINK
"http://www.speag.com/measurement/phantomsnliquids/eli4.php"
http://www.speag.com/measurement/phantomsnliquids/eli4.php
[46] (2010, May) Seibersdorf Laboratories. [Online]. HYPERLINK
"http://www.seibersdorf-laboratories.at/" http://www.seibersdorf-laboratories.at/
[47] "Human exposure to radio frequency fields from hand-held and body-mounted
wireless communication devices - Human models, instrumentation, and procedures
- Part 1: Procedure to determine SAR in close proximity to the ear (300 MHz to 3
GHz)," IEC, standard 62209-1, 2005.
[48] J. S. McLean, "A re-examination of the fundamental limits on the radiation Q of
electrically small antennas," IEEE Trans. Antennas Propagat., vol. 44, no. 5, pp.
672-676, May 1996.
[49] J. Guterman, Y. Rahmat-Samii, A.A. Moreira, C. Peixeiro, "Radiation from
Commercially Viable Antennas for PCMCIA Cards Housed in Laptops," in Mobile
and Wireless Communications Summit, July 2007, pp. 1-4.
[50] M. Ali, M.G. Douglas, A.T.M Sayem, A. Faraone, and C.-K. Chou, "Threshold
power of canonical antennas for inducing SAR at compliance limits in the 300–
3000 MHz frequency range," IEEE Trans. Electromagn. Compat., vol. 49, no. 1,
pp. 143-152, Feb. 2007.
[51] "IEEE Recommended Practice for Determining the Peak Spatial-Average Specific
88
Absorption Rate (SAR) in the Human Head from Wireless Communications
Devices: Measurement Techniques," IEEE Standards Coordinating Committee 34,
New York, IEEE Standard 1528-2003, Dec. 2003.
[52] B. Beard, W. Kainz, T. Onishi, T. Iyama, S. Watanabe, O. Fujiwara, J. Wang, G.
Bit-Babik, A. Faraone, J. Wiart, A. Christ, N. Kuster, A-K. Lee, H. Kroeze, M.
Siegbahn, J. Keshvari, H. Abrishamkar, W. Simon, D. Manteuffel, and N.
Nikoloski, "Comparisons of computed mobile phone induced SAR in the SAM
phantom to that in anatomically correct models of the human head," IEEE Trans.
Electromagn. Compat., vol. 48, no. 2, pp. 397-407, May 2006.
[53] W. Kainz, A. Christ, T. Kellom, S. Seidman, N. Nikoloski, B. Beard, and N.
Kuster, "Dosimetric comparison of the specific anthropomorphic mannequin
(SAM) to 14 anatomical head models using a novel definition for the mobile phone
positioning," Phys. Med. Biol., vol. 50, pp. 3423-3445, Jul 2005.
[54] A. Hirata, K. Shirai, and O. Fujiwara, "On averaging mass of SAR correlation with
temperature elevation due to a dipole antenna," Progress In Electromag. Research,
pp. 221-237, 2008.
89
Документ
Категория
Без категории
Просмотров
0
Размер файла
3 554 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа